Do you remember Captain James Kirk using his wrist watch to communicate with the crew of the Starship Enterprise back in 1966?

Today, almost after 50 years, it has finally become a reality!

Digital disruption is occurring in all business functions all around the world. Wearables are becoming mainstream and disrupting almost every industry, with the biggest impact being seen in customer service, healthcare and manufacturing.

Wearables currently stand at the stage where smartphones were back in 2007. Apple had just launched iPhone and the App store, but nobody could envision the vast range of applications that would soon become available. At that point, the iPhone was just considered to be a better phone, a music repository and a way to browse the web.
That is now a thing of the past.

Today, wearables come in various forms, like smart watches, health trackers, Google Glass, interactive clothing, gesture controllers and list goes on. With wearables, we can enter into an exciting new realm of augmented reality, with an enhanced experience of what we see, hear and touch.

Insurers are using wearables like Google Glass to record claims information in the field in order to process them faster.
In healthcare, implanted bio sensors can capture and transmit health data, from heart rate and blood oxygen levels, to glucose sensors — to help identify risks and make diagnoses.

Smart watches can alert users when their blood sugar is low or if they have an irregular heartbeat. With wearables, anyone can carry a personal trainer on their body at all times.

Here are some other examples of wearables in our lives today:

ADAMM is a wearable technology that provides a complete solution for managing your asthma. It collects data on cough counting, respiration and heart rate, along with medication reminders provided in an app or online portal for accessing your daily status from anywhere.

If you or someone you know suffers from lower back pain, Valedo may be a solution. The device attaches to a person’s back and uses smart sensors that communicate with a companion app to guide the user through a series of therapeutic exercises.

Helius by Proteus Digital Health is the first-ever digestible microchip, and is used to detect when a patient takes their medication. The data is transmitted to a companion app, enabling doctors and caregivers to tell if the person is taking their prescribed medicines at the correct time.

Physicians at Indiana University Health Methodist Hospital used Google Glass to perform a surgery to help remove a tumor and reconstruct an abdominal wall. Some hospitals are hoping to improve training by using wearable cameras to stream and record live surgeries as seen through the eyes of a surgeon.

Disney has developed the wearable MagicBand, a wrist band capable of monitoring visitors and collecting data about their behavior in theme parks. It also enables visitors to pay for food or merchandise, access hotel rooms, manage tickets, and skip the lines at popular attractions.

Google’s smart contact lens prototype helps measure blood glucose levels in tears for people with diabetes.

Nymi is a biometric security wristband that could someday replace all your passwords and keys.
These wearables have the potential to make our lives healthy, more secure, and more convenient — but there are a number of challenges which need to be resolved soon.

Sooner or later, all of us are identified by the data we generate, and wearables represent a quantum leap in the type and quantity of data collected — which is both an interesting and a scary proposition. 

As wearables become more mainstream, consumers need to be aware of what data is shared, which third parties have access to it, and what they will do with that information.

Originally posted on Data Science Central

Guest blog post by Bernard Marr

Apple’s old slogan was “Think Different” – and while it is now retired, and the ethos may not be as apparent in the company’s products as it once was, it is true for their approach to Big Data.

In some ways, despite being the most profitable tech company in the world, Apple found itself having to play catch-up with Big Data.

While Apple traditionally employed teams of highly paid experts in aesthetics and design to produce systems that they thought people would want to use, competitors like Google examined user data to see how people actually were using them.

This gave those competitors an edge with the everyday apps that made smart phones so popular – maps, navigation, voice recognition and other aspects of computing that we want to do on the move.

But while they may have been slow off the starting block, they have now entered the race with a strong stride. Their powerful presence in the mobile market has put their devices in the hands of millions and they have been keen to encourage development of apps that are based on monitoring and sharing of user data. A notable example is their recently announced partnership with IBM to facilitate the development of health-related mobile apps.

It has also provided a range of applications targeted at other industries, including air travel, banking and insurance, also developed in partnership with IBM and aimed at bringing analytical capabilities to users of its mobile devices in those fields.

The launch of the Apple Watch could potentially accelerate this process in a dramatic fashion – if, as many commentators are saying is possible, it turns out to be the device which finally brings wearables into the mainstream. Designed to be worn all day long, and to collect a wider variety of data thanks to additional sensors, even more personal data is available for analysis.

As well as positioning itself as an “enabler” of Big Data in other people’s lives, it has also been put to use in its own internal systems. Apple has often been secretive about the processes behind its traditionally greatest strength – product design. However it is known that Big Data also plays a part here. Data is collected about how, when and where its products – Smart phones, tablets, computers and now watches – are used, to determine what new features should be added, or how the way they are operated can be tweaked to provide the most comfortable and logical user experience.

The Siri voice recognition features of iDevices have proved popular with users too, and this is also powered by Big Data. Voice data captured by the machine is uploaded to its cloud analytics platforms, which compare them alongside millions of other user-entered commands to help it become better at recognizing speech patterns (machine learning) and more accurately match users to the data they are seeking. Apple keeps this data for two years – disassociated from your real identity and assigned with a unique anonymous indicator, as a concession to ensuring privacy. 

Like its biggest competitors, Apple also offers cloud-based storage, computing and productivity solutions, for both consumer and business use. Last month it was reported that it had purchased FoundationDB, a popular proprietary database architecture widely used for Big Data applications.  It is thought that this could be used to bring increased analytical prowess across its suite of online services such as iCloud, Apple Productivity Works (formerly iWork) and its upcoming streaming music service.

Aiming to capture a share of the market dominated by Pandora, Spotify and Google Music, this service will be built on the technology acquired by their purchase last year of Beats Music. Beats developed algorithms designed to match users with music they are likely to enjoy listening to, in a similar way to recommendation engines used by Amazon and Netflix. Sales through Apple’s iTunes service have declined as the popularity of streaming services has usurped downloading as the favorite method of accessing music online. The new service, expected in June, is Apple’s attempt to get a slice of this action.

Apple may have been slower in its uptake of Big Data and analytics than many of its rivals, but it has clearly seen that it has to play a big part in its future if it wants to stay ahead of the pack. It seems likely that it will try and use it to move away from relying on hugely expensive, episodic product releases to drive its growth as a business, and towards the more organic, constantly-regenerating model of growth favoured by its competitors in the software and services markets. If Apple can meld its trademark excellence in design and user-friendliness with innovative uses of Big Data analytics, it could continue to surprise us with products and services which become culturally ingrained in everyday life, just like the iMac, iPod and iPhone – ensuring it remains the world’s most valuable brand for some time to come.

I hope you found this post interesting. I am always keen to hear your views on the topic and invite you to comment with any thoughts you might have.

About : Bernard Marr is a globally recognized expert in analytics and big data. He helps companies manage, measure, analyze and improve performance using data.

His new book is: Big Data: Using Smart Big Data, Analytics and Metrics To Make Better Decisions and Improve Performance You can read a free sample chapter here.

Follow us @IoTCtrl | Join our Community

Summary:  This is my favorite IoT story. We are so used to IoT platforms being physical objects that we forget about the potential for biologics.  In terms of direct economic reward little will compare to this story about the IoT and cows.

This is my favorite IoT story which I first heard from Joseph Sirosh, CVP of Machine Learning for Microsoft at the spring Strata convention in San Jose.  We are so used to IoT platforms being physical objects like cars or thermostats or gaming consoles that we forget about the potential for biologics.  Of course FitBit will immediately come to mind for human beings but in terms of direct economic reward little will compare to this story about the IoT and cows.

The setting is Japan but it could just as easily have been Iowa.  The players are Fujitsu who developed this system and Microsoft Azure providing the NoSQL DB on which it runs.

The opportunity requires seeing a dairy farm as a simple manufacturing environment with cows and feed coming in one end and milk products coming out the other.  The immediate problem for the farmer is replacing or adding to his production capacity, the cows.

Dairy cattle are largely ‘produced’ by artificial insemination and we’ve been doing that long enough that we can get a 70% conception rate, but only if the procedure occurs when the cow is in estrus.  That proves much harder to predict since it mostly relies on the farmer’s experience-based intuition which turns out isn’t all that good.  In fact, farmers only get it right statistically about 55% of the time meaning that the true pregnancy rate is only 39%.

So Fujitsu reasoned if estrus detection could be made perhaps 95% accurate the pregnancy rate would go up to 67% or a whopping 70% improvement in performance.  But estrus detection isn’t all that easy since Bessie is only in estrus about once every 21 days; estrus lasts only 12 to 18 hours, and just to make it that much more difficult, usually occurs between 10 pm and 8 am, just when the poor exhausted farmer is asleep.

So first let’s talk about how this all turned out.  In the chart below the green line represents the number of steps normally taken by a cow and the yellow line represents the steps taken by a cow in estrus.  It turns out that the onset of estrus can be detected because Bessie starts doing a little dance.  The optimum time for artificial insemination is 16 hours later.  And as a very valuable added bonus it turns out that AI performed in the first two hours of the window has a much higher probability of producing a female which of course is what you want for a milk cow.

Fujitsu equipped the herd with battery powered pedometers to detect the number of steps, wirelessly transmits that information to a MS Azure DB where it’s analyzed, and notifies the farmer via mobile phone app when the time is right.  Pretty remarkable.

In field testing, over a group of 11 dairy farms, the farmers achieved annual increases in herd size, when compared with historical methods that averaged 12% and ranged upwards to 31%.

If you’re a farmer that’s all you really need to know.  But if you’re a data scientist it’s the part of the story that isn’t told above where the magic happens.

When conceiving this idea that estrus can be detected by some combination of measurable biologic factors that the cow will demonstrate, as data scientist we can only imagine the amount of poking and prodding the test cows must have endured as Fujitsu built up its data base.  I imagine there must have been multiple body temperature, blood chemistry, waste and urine chemistry, and who knows what else that was originally fed into the hopper to build the model.  Remember that IoT is a classic two step data science process.  Gather and analyze data and use classical predictive modeling techniques to find a signal in the data that predicts the desired outcome, estrus.  I imagine the data scientist at Fujitsu must have been overjoyed to find something as simple as the number of steps since it could have been much messier.

The data which is found to be predictive, the steps, is then gathered in the production environment via the pedometers and scored via the predictive models to signal the farmer that Bessie is ready to get busy.

As a bonus, Fujitsu claims that it can also detect as many as ten different diseases using the same technique.  If you’d like to see Joseph Sirosh tell the story it’s available on YouTube at https://www.youtube.com/watch?v=oY0mxwySaSo.

October 27, 2015

Bill Vorhies, President & Chief Data Scientist – Data-Magnum - © 2015, all rights reserved.

About the author:  Bill Vorhies is President & Chief Data Scientist at Data-Magnum and has practiced as a data scientist and commercial predictive modeler since 2001.  Bill is also Editorial Director for Data Science Central.  He can be reached at:

Bill@Data-Magnum.com or Bill@datasciencecentral.com

Guest blog post by ajit jaokar

The Open Cloud – Apps in the Cloud 

Smart Data

Based on my discussions at Messe Hannover , this blog explores the potential of applying Data Science to manufacturing and process control industries. In my new course at Oxford University (Data Science for IoT) and community (Data Science and Internet of Things ), I explore application of predictive algorithms to Internet of Things (IoT) datasets. 

The Internet of Things plays a key role here because sensors in machines and process control industries generate a lot of data. This data has real, actionable business value (Smart Data). The objective of Smart data is to improve productivity through digitization. I had a chance to speak to Siemens management and engineers about how this vision of Smart Data is translated into reality

When I discussed the idea of Smart Data with Siegfried Russwurm, Prof. Dr.-Ing. - Member of the Managing Board of Siemens AG ,  he spoke of key use cases that involve transforming big data into business value by providing context, increasing efficiency  and addressing large, complex problems. These include applications for Oil rigs, wind turbines and process control industries etc. In these industries, the smallest productivity increase translates to huge commercial gains.  

This blog is my view on how this vision (Smart data) could translate into reality within the context Data Science and IoT.

Data: the main driver for Industrie 4.0 ecosystem

 At Messe  Hannover, it was hard to escape the term ‘Industry 4.0’ (in German – Industrie 4.0). Broadly, Industry 4.0 refers to the use of electronics and IT to automate production and to create intelligent networks along the entire value chain that can control each other autonomously. Machines generate a lot of Data. In many cases, if you consider the large installation such as an Oil Rig, this data is bigger than the traditional ‘Big Data’.  Its use case is also slightly different i.e. the value does not like in capturing a lot of data from outside the enterprise – but rather in capturing (and making innovative uses of) a large volume of data generated within the enterprise.  The ‘smart’ in smart data is predictive and algorithmic. Thus, Data is the main driver of Industry 4.0 and it’s important to understand the flow of Data before it can be optimized

The flow of Data in the Digital Enterprise

The ‘Digital factory’ is already a reality. For instance,  Industrial Ethernet standards like Profinet, PLM(Product lifecycle management) software like Teamcenter  and Data models for lifecycle engineering and plan management such as Comos. To extend the Digital factory  to achieve end-to-end interconnection and autonomous operation across the value chain (as is the vision of Industry 4.0), we need a component  in the architecture.  

The Open Cloud: Paving the way for Smart Data analytics

In that context,  the cooperation of Siemens with SAP to create open cloud platform. Is very interesting. The Open Cloud enables ‘apps in the cloud’  based on the intelligent use of large quantities of data. The SAP Hana architecture based on in-memory, columnar database provides analytics services in the Cloud. For instance, the "Asset Analytics"(to increase the availability of machines through online monitoring, pattern recognition, simulation,  prediction of issues) and  “Energy Analytics" ( revealing hidden energy savings potential)

Conclusions

While it is early days, based on the above, the manufacturing domain offers real value and tangible benefits to customers. Even now, we see the customers  who harness value from large quantities of Data through predictive analytics stand to gain significantly. I will cover this subject in more detail as it evolves. 

About the author

Ajit''s work spans research, entrepreneurship and academia relating to IoT, predictive analytics and Mobility. His current research focus is on applying data science algorithms to IoT applications. This includes Time series, sensor fusion and deep learning.  This research underpins his teaching at Oxford University (Big Data and Telecoms) and the City sciences program at the Technical University of Madrid (UPM). Ajit also runs a community/learning program through his company - futuretext for Data Science and IoT

Follow us @IoTCtrl | Join our Community

Here are a few thoughts from @dataguild on IoT as applied to Data Science. Thanks to @MacSlocum, @JonBruner and the @OReillySolid crew for a great show in San Francisco last month. 

Guest blog post by Ajit Jaokar

In this post, I propose that IoT analytics should be a part of 'Smart objects' and discuss the implications of doing so

The term ‘Smart objects’ has been around from the times of Ubiquitous Computing.

However, as we have started building Smart objects, I believe that the meaning and definition has evolved.

Here is my view on how the definition of Smart Objects has changed in the world of Edge Computing and increasing processing capacity

At a minimum, a smart Object should have 3 things

a) An Identity ex ipv6
b) Sensors / actuators
c) A radio (Bluetooth / cellular etc)

In addition, a smart object could incorporate

a) Physical context ex location
b) Social context ex proximity in social media

To extend even more, Smartness could incorporate analytics

Some of these analytics could be performed on the device itself ex computing at the edge concept from Intel, Cisco and others.

However, Edge Computing as discussed today, still has some limitations

For example:

a)     The need to incorporate multiple feeds from different sensors to reach a decision ‘at the edge’

b)    The need for a workflow process i.e. actions based on readings – again often at the edge with it’s accompanying security and safety measures

To manage multiple sensor feeds, we need to understand concepts like sensor fusion (pdf) (source freescale).

We already have some rudimentary workflow through mechanisms like IFTTT(If this then that)

In addition, the rise of CPU capacity leads to greater intelligence on the device – for example Qualcomm Zeroth platform which enables Deep learning algorithms on the device.

So, in a nutshell, its a evolving concept especially if we include IoT analytics in the definition of Smart objects (and that some of these analytics could be performed at the Edge)  ..

We cover these ideas in the #DataScience for #IoT course and also at the courses I teach at Oxford University

Comments welcome

Follow us @IoTCtrl | Join our Community

Guest blog post by Ajit Jaokar

Time Series IoT applications in Railroads

Authors: Vinay Mehendiratta, PhD, Director of Research and Analytics at Eka Software

and Ajit Jaokar, Data Science for IoT course  

This blog post is part of a series of blogs exploring Time Series data and IoT.

The content and approach are part of the Data Science for Internet of Things practitioners course.  

Please contact info@futuretext.com for more details.

Only for this month, we have a special part-payment pricing for the course (which begins in November).

We plan to develop these ideas more – including an IoT toolkit in the R programming language for IoT datasets. You can sign up for more posts from us HERE

Introduction 

Over the last fifteen years, Railroads in the US, Europe and other countries have been using  RFID devices on their locomotives and railcars.  Typically, this Information is stored in traditional (i.e. mostly relational) databases. Information from the RFID scanner provides information about the railcar number and locomotive number. This railcar number is then mapped to existing railcar and train schedule. Timestamp information on scanned data also provides us the sequence of cars on that train. Information from data obtained by scanning RFID on locomotive provides us the number of locomotives and the total horsepower assigned to the train. It also informs whether locomotive is coupled in front of the train or rear of the train.

The scanned data  requires cleansing. Often, readings  from a railcar RFID are  missing at certain scanner. In this case, the missing value is estimated by looking at the scanner reading  before and after the problematic scanner to estimate the time of arrival.

Major Railroads have also defined their territory using links where a link is the directional connection between two nodes.  Railroads have put RFID scanners at major links. 

 An RFID gives information on railcar sequence in train, locomotive consist, and track in real-time.  Railroads store this real-time and historical data for analysis.

Figure 1: Use cases of Rail Time Series Data

Figure 1 above shows use cases of time series data in railroad industry. We believe that all of these use cases are applicable for freight railroads. These use cases can also be used for passenger railroads with some changes.  They involve the use of Analytics and RFID

Uses of Real-Time Time Series Data

Here are some ways that time series data is/can be used in railroads in real-time.

1. Dispatching: Scanner data is being used for dispatching decisions for many years now.  Scanner data is used to display the latest location of trains. Dispatchers use this information, track type, train type, time table information to determine the priority that should be assigned to various trains.
2. Information for Passengers: Passengers can use train arrival and departure estimates for planning their journey.

Uses of Historical Time Series Data:

Here are some ways that historical time series data is/can be used in railroads.

• Schedule Adherence Identify trains that are consistently delayed: We can identify trains that are on Schedule, delayed or earlier. . We can identify trains that consistently occupy tracks more than the schedule permit. These are the trains that should be considered for a schedule change. These are the trains that are candidate for root cause analysis.

• Better Planning: We would be able to determine if planned ‘sectional running time’ are accurate or need to be checked. Sectional run times are generally determined based on experience and are estimates at network level but don’t consider local infrastructure (signal, track type). Sectional running time is used in development of train schedule and maintenance schedule at network and local level

• Infrastructure Improvemen - Track Utilization: We can identify the section of track where trains have the highest occupancy. This would lead us to identify tracks that are being operated near track capacity or above track capacity. Assumption here is that Utilization above track capacity would result in delays. We can identify the set of trains, tracks, time of day, day of the week when occupancy is high and low. This would provide us insights in train movement and perhaps provide suggestions on train schedule change. We might be able to determine if trains are held up at station/yards or on mainline.  An in-depth and careful analysis can help us determine if attention needs to be paid to yard operations or mainline operations.

• Simulation Studies: RFID scan data provides us actual time of arrival and departure for every car (hence train). Modelers do create hypothetical trains to feed to simulation studies. This information (actual train arrival/departure time at every scanner, train consist, locomotive consist) is used in infrastructure expansion projects.

• Maintenance Planning : Historical Occupancy of tracks would enable us to identify time windows when maintenance should be scheduled in future. Railroads use inspection cars to inspect and record track condition regularly. Some railroads are facing the challenge of getting the right geo coordinates for segment of track. Careful insights of this geo and time series data measure track health and deterioration. Satellite imagery data is becoming available frequently. A combination of these two sources can do well to inspect tracks, schedule maintenance, predict track failures, and move maintenance gangs.

• Statistical Analysis of Railroad Behavior

1. We can map train behavior with train definition (train type, schedule, train speed, train length) and track definition (signal type, track class, grade, curve, authority type) and identify patterns.
2. Passenger trains do affect the operations of freight trains. Scanner data can be used to determine the delay imposed on freight trains
3. Time series information of railcars can be used to identify misrouted cars or lost cars.
4. Locomotive consist information and time series data based performance can be used together to determine the best locomotive consist such as make, horsepower (historically) for every track segment
5. Locomotive is a costly asset for any railroad. Time series data can easily be used to determine locomotive utilization.

•  Demand Forecasting : Demand for railroad empty cars is known as an indicator of a country’s economy. While demand of railroad cars vary with car type and macro-economic factors, it is worth making efforts getting insights on historical perspective. Number of cars by car type can be estimated and forecasted for every major origin-destination pair. Number of train starts and train ends at every origin and destination can be used to forecast the number of trains for a future month. Number of trains forecasted would help a railroad determine the number of crew, locomotives. It would also help railroad determine the load that tracks would go through.  Number of forecasted trains can be used in infrastructure studies.

• Safety: Safety is  the most important feature of railroad culture. Track maintenance, track wear and tear ( track utilization) are all related to safety. Time series data of railcars, signal type, track type, train type, accident type, train schedule can all be analyzed together to identify potential relationship (if any) between various relevant factors.

• Train Performance Calculations: What is the unopposed running speed on a track with a given grade, curve, locomotive consist, car type, wind direction and speed?  These factors were  determined by Davis [1] in 1926. Could time series data help us calibrate the co-efficient of Davis’s equation for railcars with new design?

• Planning and Optimization: All findings above can be used to develop smarter optimization models for train schedule, maintenance planning, locomotive planning, crew scheduling, and railcar assignment.

Conclusion:

In this article,  we have highlighted some use cases of time series data for Railroads. There are many more factors that could be considered especially in the use of Technology for implementing these Time series algorithms. In subsequent sections, we will show how some of these use cases could be implemented based on the R programming language.

To know more about the Data Science for Internet of Things practitioners course.  Please contact info@futuretext.com for more details. You can sign up for more posts from us HERE

Reference:

1. Davis, W.J, Jr.: The tractive resistance of electric locomotives and cars, General Electric Rewiew, vol. 29, October 1926. 

Follow us @IoTCtrl | Join our Community

At IoT Central we aim to cover all things industrial and IoT. Our site is segmented into five channels: Platforms, Apps & Tools, Data, Security and Case Studies. If you’re going to connect everything in the world to the Internet you should expect to cover a lot. That means plenty of reading, sharing and discussing.  

To tackle the reading part we reached out to our peers and friends and put together the 10 best books to read now on IoT. From theoretical to technical, we tried to find the most important and current reading while throwing in one or two relevant classics.

Below is the list we compiled. What books would you recommend?

Shaping Things

By Bruce Sterling

I first came across Bruce Sterling’s name when he wrote the November 1996 Wired cover story on Burning Man. I happened to attend the desert arts festival for the first time that year and Bruce’s prose nailed the the experience. I’ve been a fan of his ever since. "Shaping Things is about created objects and the environment, which is to say, it's about everything," says Bruce. This is a great higher level book that looks at the technosocial transformation needed to understand our relationship between the Internet of Things and the environment in which it exists.

The Hardware Startup

By Renee DiResta, Brady Forrest, Ryan Vinyard

Consumer Internet startups seem to get all the media ink these days - think AirBnB, Instagram, What’sApp, Uber. But many forget that much of the technological innovation began with hardware - think Fairchild Semiconductor, Xerox PARC and the stuff that came out of IBM. With an emphasis on ‘Things,’ IoT is set to usher in a new era of hardware startups and any entrepreneur in this space should find this book to be a valuable read.

IoT M2M Cookbook

By Harald Nauman

If IoT devices can’t communicate, you’re not going to get much use out of them. Someone pointed me to Harald Naumann’s book IOT/M2M Cookbook. Harold is an M2M evangelist with a primary interest in implementation of wireless applications. His blog is chocked full of technical tips on wireless communications.

IoT Central members can see the full list here. Become a member today here. 

guest blog by Jin Kim, VP Product Development for Objectivity, Inc.

Almost any popular, fast-growing market experiences at least a bit of confusion around terminology. Multiple firms are frantically competing to insert their own “marketectures,” branding, and colloquialisms into the conversation with the hope their verbiage will come out on top.

Add in the inherent complexity at the intersection of Business Intelligence and Big Data, and it’s easy to understand how difficult it is to discern one competitive claim from another. Everyone and their strategic partner is focused on “leveraging data to glean actionable insights that will improve your business.” Unfortunately, the process involved in achieving this goal is complex, multi-layered, and very different from application to application depending on the type of data involved.

For our purposes, let’s compare and contrast two terms that are starting to be used interchangeably – Information Fusion and Data Integration. These two terms in fact refer to distinctly separate functions with different attributes. By putting them side-by-side, we can showcase their differences and help practitioners understand when to use each.

Before we delve into their differences, let’s take a look at their most striking similarity. Both of these technologies and best practices are designed to integrate and organize data coming in from multiple sources in order to present a unified view of data for consumption by various applications to derive actionable insights, thus making it easier for analytics applications to use and derive the “actionable insights” everyone is looking to generate.

However, Information Fusion diverges from Data Integration in a few key ways that make it much more appropriate for many of today’s environments.

• Data Reduction – Information Fusion is, first and foremost, designed to enable data abstraction. So, while data integration focuses on combining data to create consumable data, Information Fusion frequently involves “fusing” data at different abstraction levels and differing levels of uncertainty to support a more narrow set of application workloads.

• Handling Streaming/Real-Time Data – Data Integration is best used with data-at-rest or batch-oriented data. The problem is that the most compelling applications associated with Big Data and the Industrial Internet of Things are often based on streaming, sensor data. Information Fusion is capable of integrating, transforming and organizing all manner of data (structured, semi-structured, and unstructured), but specifically time-series data, for use by today’s most demanding analytics applications to bridge the gap between Fast Data and Big Data. Another way to put this is Data integration creates an integrated set of data where the larger set is retained. By comparison, Information Fusion uses multiple techniques to reduce the amount of stateless data and provide only the stateful, valuable and relevant, data to deliver improved confidence.

• Human Interfaces – Information Fusion also adds in the opportunity for a human analyst to incorporate their own contributions to the data in order to further reduce uncertainty. By adding and saving inferences and detail that can only be derived with human analysis and support into existing and new data, organizations are able to maximize their analytics efforts and deliver a more complete “Big Picture” view of a situation.

As you can see, Information Fusion, unlike Data Integration, focuses on deriving insight from real-time streaming data and enriching this stream with semantic context from other Big Data sources. This is a critical distinction, as todays most advanced, mission-critical, analytical applications start looking to Information Fusion to add real-time value.

Originally posted on Data Science Central

Follow us @IoTCtrl | Join our Community

Guest blog post by Venkat Viswanathan

More than 10 years ago, on an early summer afternoon in 2005, I recall an interesting conversation with a friend about the potential of analytics, in an empty coffee shop on the beach. Having spent much of the previous decade helping clients derive value from traditional Business Intelligence (BI) and other IT implementations, we were conjecturing that the next wave might be business teams enhancing their decision making with better data and more detailed analysis, sometimes adopting advanced math. We spent hours talking about our collective experience, and the more we spoke, the more it piqued my curiosity.

Deriving Actionable Insights

Around that time, the emphasis on analytics was primarily for its predictive capabilities. Mainstream media played this up with cover stories on BusinessWeek (“Math will Rock Your World”, Jan 2005, ), by Stephen Baker) and in popular business books like Freakonomics, Steven Levitt & Co., 2005 and later Super Crunchers, Ian Ayres, 2007. However, within a couple of years, it became evident that the vast majority of businesses were still not convinced about the quality of data they collect, (many still aren’t!) and the fundamental need was to create organizational change. Companies needed to understand there was a better way to derive actionable business insights that would result in better decisions on the ground. More than just help with solving tough math problems, large enterprises needed the ability to dig deep into the data, working in the trenches to create a center of excellence tailored to address their specific gaps in knowledge, talent availability and problem solving skills essential to realize the analytics potential.

Along the way, changes in the business environment, innovations in the technology ecosystem, and the successes of data driven business models – creating digital role models like Google AdWords, and Facebook early on and Uber and AirBnB more recently – have redefined the Analytics opportunities for businesses. As social media platforms took off, companies took notice, and around 2009, consumer brands started warming up to “social listening.” Decoding social media conversations and other unstructured data became an important part of their analytics needs. Knowledge of semantic analysis, platform APIs, and client business context proved crucial in combining machine intelligence with human intelligence and delivering digital business insights.

Visualizing the Future

Another wave of Analytics for enterprises was around data visualization. I love saying “visualization is to analytics what email was to the Internet.” A simple, engaging and dynamic application that is used every day and helps with managing a business. The pioneering work and books of Edward Tufte and Stephen Few were an inspiration for many as they realized both the power of data visualization and that they needed help to do it well. Tableau’s meteoric rise and ever expanding footprint meant that by 2012 many companies had already invested in it and were looking for knowledge and expertise in embedding it within their businesses. The visual medium is crucial to democratizing access to data and allowing business users to quickly navigate to actionable insights.

Fast forward to 2015, and peering ahead, the future is bright! With the advent of multiple iterations of drones, intelligent machines, connected devices and wearables – from watches to activity trackers to devices that infer your state of mind from your breath, and the increasing buzz around the potential for the Internet of Things (IoT), we are at the cusp of some fundamental breakthroughs. In less than a decade, our machine generated data and algorithm driven interactions will far outnumber traditional data sources and applications. Analytics will gain all new interpretations, and business applications. The future of the space has infinite possibilities and is exceedingly exciting. As we head toward the close of 2015, we can look forward to the coming year assured that it will be filled with innovation and advancement. I feel fortunate to be a part of this incredible industry and look forward to the many advancements to come.

Follow us @IoTCtrl | Join our Community

Guest blog post by Ajit Jaokar

Introduction

 This blog is a review of two books. Both are available for free from the MapR site, written by Ted Dunning and Ellen Friedman (published by O Reilly) : About Time Series Databases: New ways to store and access data andA new look at Anomaly Detection

 The  MapR platform is a key part of the Data Science for the Internet of Things (IoT) course – University of Oxford and I shall be covering these issues in my course

 In this post, I discuss the significance of Time series databases from an IoT perspective based on my review of these books. Specifically, we discuss Classification and Anomaly detection which often go together for typical IoT applications. The books are easy to read with analogies like HAL (Space Odyssey ) and I recommend them.

Time Series data

The idea of time series data is not new. Historically, time series data can be stored even in simple structures like flat files. The difference now is the huge volume of data and the future applications possible by collecting this data – especially for IoT. These large scale time series databases and applications are the focus of the book. Large scale time series applications typically need a NoSQL database like Apache Cassandra, Apache HBase,  MapR-DB etc.  The book’s focus is Apache HBase and MapR-DB for the collection, storage and access of large-scale time series data.

  Essentially, time series data involves measurements or observations of events as a function of the time at which they occurred. The airline ‘black box’ is a good example of a time series data. The black box records data many times per second for dozens of parameters throughout the flight including altitude, flight path, engine temperature and power, indicated air speed, fuel consumption, and control settings. Each measurement includes the time it was made. The analogy applies to sensor data. Increasingly, with the proliferation of IoT, Time series data is becoming more common and universal. The data so acquired through sensors is typically stored in Time Series Databases.  The TSDB (Time series database) is optimized for best performance for queries based on a range of time

Time series data applications

Time series databases apply to many IoT use cases for example:

• Trucking, to reduce taxes according to how much trucks drive on public roads (which sometimes incur a tax). It’s not just a matter of how many miles a truck drives but rather which miles.
• A smart pallet can be a source of time series data that might record events of interest such as when the pallet was filled with goods, when it was loaded or unloaded from a truck, when it was transferred into storage in a warehouse, or even the environmental parameters involved, such as temperature.
• Similarly, commercial waste containers, called dumpsters in the US, could be equipped with sensors to report on how full they are at different points in time.
• Cell tower traffic can also be modelled as a time series and anomalies like flash crowd events that can be used to provide early warning.
• Data Center Monitoring can be modelled as a Time series to predict  outages, plan upgrades
• Similarly, Satellites, Robots and many more devices can be modelled as Time series data

From these readings captured in a Time Series database, we can derive analytics such as:

Prognosis: What are the short- and long-term trends for some measurement or ensemble of measurements?

Introspection: How do several measurements correlate over a period of time?

Prediction:  How do I build a machine-learning model based on the temporal behaviour of many measurements correlated to externally known facts?

Introspection:  Have similar patterns of measurements preceded similar events?

Diagnosis:  What measurements might indicate the cause of some event, such as a failure?

Classification and Anomaly detection for IoT

The books gives examples of usage of Anomaly detection and Classification for IoT data.

For Time series IoT based readings, anomaly detection and Classification go together. Anomaly detection determines what normal looks like, and how to detect deviations from normal.

When searching for anomalies, we don’t know what their characteristics will be in advance. Once we know characteristics, we can use a different form of machine learning i.e. classification

Anomaly in this context just means different than expected—it does not refer to desirable or un‐ desirable. Anomaly detection is a discovery process to help you figure out what is going on and what you need to look for. The anomaly-detection program must discover interesting patterns or connections in the data itself.

Anomaly detection and classification go together when it comes to finding a solution to real-world problems. Anomaly detection is used first in the discovery phase—to help you figure out what is going on and what you need to look for. You could use the anomaly-detection model to spot outliers, then set up an efficient classification model to assign new examples to the categories you’ve already identified. You then update the anomaly detector to consider these new examples as normal and repeat the process

The book goes on to give examples of usage of these techniques in EKG

For example, for the challenge of finding an approachable, practical way to model normal for a very complicated curve such as the EKG, we could use a type of machine learning known as deep learning.

 Deep learning involves letting a system learn in several layers, in order to deal with large and complicated problems in approachable steps. Curves such as the EKG have repeated components separated in time rather than superposed. We take advantage of the repetitive and separated nature of an EKG curve in order to accurately model its complicated shape to detect normal patterns using Deep learning

The book also refers to a Data structure called t-Digest for Accurate Calculation of Extreme Quantiles  t-digest was developed by one of the authors, Ted Dunning, as a way to accurately estimate extreme quantiles for very large data sets with limited memory use. This capability makes t-digest particularly useful for selecting a good threshold for anomaly detection. The t-digest algorithm is available in Apache Mahout as part of the Mahout math library. It’s also available as open source athttps://github.com/tdunning/t-digest

Anomaly detection is a complex field and needs a lot of data.

For example: what happens if you only save a month of sensor data at a time, but the critical events leading up to a catastrophic part failure happened six weeks or more before the event?

IoT from a large scale Data standpoint

To conclude, much of the complexity for IoT analytics comes from the management of Large scale data.

Collectively, Interconnected Objects and the data they share make up the Internet of Things (IoT).

Relationships between objects and people, between objects and other objects, conditions in the present, and histories of their condition over time can be monitored and stored for future analysis, but doing so is quite a challenge.

However, the rewards are also potentially enormous. That’s where machine learning and anomaly detection can provide a huge benefit.

For Time series, the book covers themes such as

Storing and Processing Time Series Data

The Direct Blob Insertion Design

Why Relational Databases Aren’t Quite Right

Architecture of Open TSDB

Value Added: Direct Blob Loading for High Performance

Using SQL-on-Hadoop Tools

Using Apache Spark SQL

 Advanced Topics for Time Series Databases(Stationary Data, Wandering Sources, Space-Filling Curves )

For Anomaly detection:

Windows and Clusters

 Anomalies in Sporadic Events

Website Traffic Prediction

Extreme Seasonality Effects

Etc

Links again:

About Time Series Databases: New ways to store and access data and 

A new look at Anomaly Detection  by Ted Dunning and Ellen Friedman (published by O Reilly).

Also the link for Data Science for the Internet of Things (IoT) course – University of Oxford where I hope to cover these issues in more detail in context of  MapR

Follow us @IoTCtrl | Join our Community

Guest blog post by Bill Vorhies

Summary:  NIST weighs in on the Internet of Things to create a common vocabulary and development roadmap.

In July we wrote about the 7 volume Big Data Technology Roadmap being developed by the National Institute of Standards and Technology (NIST) which is part of the Department of Commerce.  You had an opportunity to review and comment on this final draft before publication.  See that article here.

Well they’re back and this time with a final draft of their comprehensive roadmap for the Internet of Things (IoT).  Despite the fact the IoT is the widely accepted name for this field, NIST elected to call their study “Framework for Cyber Physical Systems (CPS)”.  Doesn’t quite roll off the tongue like IoT but don’t let that deter you from taking a look.

Like its Big Data predecessor, this is the result of collaboration among business, academia, and government experts organized into the Cyber-Physical Systems Public Working Group (CPS PWG).  At 227 pages it’s a comprehensive reference of all things IoT from a wide range of contributors.

When the CPS-PWG decided on its somewhat unusual name it appears they were trying to draw a definition around a number of phrases, some of which have fallen out of use.  In addition to IoT, they include the domains of M2M (machine to machine), the industrial internet, and smart cities among others. Those of us on the Big Data and predictive analytics side of things tend to view this all as IoT.

Despite the odd naming, they created an interesting taxonomy about what makes CPS (IoT) different and distinguishable from other things:

The combination of the cyber and the physical, and their connectedness, is essential to CPS

• CPS devices may be repurposed beyond applications that were their basis of design –e.g., a cell phone in a car may be used as a mobile traffic sensor; energy usage information may be used to diagnose equipment faults.
• CPS networks may have “brokers” and other infrastructure-based devices and aggregators that are owned and managed by third parties, resulting in potential trust issues – e.g., publish and subscribe messaging, certificate authorities, type and object registries.
• CPS are noted for enabling cross-domain applications – e.g., the intersection of manufacturing and energy distribution systems, smart cities, and consumer-based sensing.
• Because CPS are designed to interact directly with the physical world, there is a more urgent need for emphasis on security, privacy, safety, reliability, and resilience, and corresponding assurance for pervasive interconnected devices and infrastructures.
• CPS should be composable and may be service based. Components are available that may be combined into a system dynamically and the system architecture may be modified during runtime to address changing concerns. There are challenges, however. For example, timing composability may be particularly difficult. Also, it may not always be necessary or desired to purchase assets to build a system; instead, services can be purchased on a per-use basis, only paying for using the resources needed for a specific application and at the specific time of usage.

The document which you can download here covers nine broad areas:

1. Functional
2. Business
3. Human
4. Trustworthiness
5. Timing
6. Data
7. Boundaries
8. Composability
9. Lifecycle

There are also excellent listings of references for those wishing a deeper dive and a good appendix of Definitions and Acronyms.  Also a number of well detailed use cases to spur the imagination of you IoT entrepreneurs.

The good news about these NIST studies, both this IoT study and its Big Data brother is that they are quite comprehensive and represent the thinking of a very wide range of public and private experts.  They are also completely public domain.  The bad news is that they take a long time to complete between their committee development and public review process.  This one started in 2014 and I don’t find much here about real time or streaming analytics or combined analytic and transactional databases like SAP HANA or VoltDB that are today’s forefront of IoT enablement.

If you want to be part of the process, the Working Group is taking public comment via email until November 2, 2015.

The template for submitting comments is available here.  Please submit comments using the spreadsheet template by November 2, 2015 via email to DraftCPSFrameworkComments@cpspwg.org.

The Draft CPS Framework is freely available for download here. An additional Technical Annex, Timing Framework for Cyber-Physical Systems, is also freely available for download here.  Their homepage is found here http://www.nist.gov/cps/cps-pwg-workshop.cfm.

September 28, 2015

Bill Vorhies, President & Chief Data Scientist – Data-Magnum - © 2015, all rights reserved.

About the author:  Bill Vorhies is President & Chief Data Scientist at Data-Magnum and has practiced as a data scientist and commercial predictive modeler since 2001.  Bill is also Editorial Director for Data Science Central.  He can be reached at:

Bill@Data-Magnum.com or Bill@DataScienceCentral.com

Follow us @IoTCtrl | Join our Community

This week at the Gartner Symposium/ITxpo 2015 upwards of 10,000 CIOs and business technology professionals from around the world are gathering to talk all things IT. Gartner regularly polls their clients and today released the Top 10 Strategic Technology Trends for 2016.

Gartner defines a strategic technology trend as one with the potential for significant impact on the organization. Factors that denote significant impact include a high potential for disruption to the business, end users or IT, the need for a major investment, or the risk of being late to adopt.

As this community would expect, IoT dominates the majority of the list.

In this latest report, everything is a device and the general idea is that the digital mesh is a dynamic network linking various endpoints.

IoT related trends include:

The Device Mesh

The device mesh refers to an expanding set of endpoints people use to access applications and information or interact with people, social communities, governments and businesses. The device mesh includes mobile devices, wearable, consumer and home electronic devices, automotive devices and environmental devices — such as sensors in the Internet of Things (IoT).

Ambient User Experience

While this trend focuses on augmented and virtual reality, IoT sensors play a key role in how this is implemented.

Information of Everything

Everything in the digital mesh produces, uses and transmits information. Advances in semantic tools such as graph databases as well as other emerging data classification and information analysis techniques will bring meaning to the often chaotic deluge of information.

Advanced Machine Learning

Gartner explores deep neural nets (DNNs), (an advanced form of machine learning particularly applicable to large, complex datasets) and claims this is what makes smart machines appear "intelligent." DNNs enable hardware- or software-based machines to learn for themselves all the features in their environment, from the finest details to broad sweeping abstract classes of content.

Autonomous Agents and Things

Gartner Research Fellow David Cearley says, "Over the next five years we will evolve to a postapp world with intelligent agents delivering dynamic and contextual actions and interfaces. IT leaders should explore how they can use autonomous things and agents to augment human activity and free people for work that only people can do. However, they must recognize that smart agents and things are a long-term phenomenon that will continually evolve and expand their uses for the next 20 years."

Adaptive Security Architecture

Security and IoT should go hand-in-hand. Gartner says that relying on perimeter defense and rule-based security is inadequate, especially as organizations exploit more cloud-based services and open APIs for customers and partners to integrate with their systems.

Advanced System Architecture, Mesh App and Service Architecture

These are three of the ten trends that I’m summarizing into one. All of these require more computing power and new ways of deploying software. Say goodbye to the monolithic approach and welcome agility. Application teams must create new modern architectures to deliver agile, flexible and dynamic cloud-based applications with agile, flexible and dynamic user experiences that span the digital mesh.

Internet of Things Platforms

IoT platforms complement the mesh app and service architecture and Mr. Cearley rounds out the trends by stating, "Any enterprise embracing the IoT will need to develop an IoT platform strategy, but incomplete competing vendor approaches will make standardization difficult through 2018.”

Lots of work to still do. Further reading here.

Guest blog post by Bernard Marr

Elite level athletes have long had the ability to integrate data analysis principles into their training – monitoring and crunching data on their performance to help them break personal bests and world records.

Thanks to the explosion of the Internet of Things – the idea that just about any everyday object can be made “smart”, and able to collect data and communicate wirelessly – these sort of insights are now available to athletes and players at any level.

Here’s a rundown of what I think are five of the best Internet of Things enabled sports and training gadgets and apps which can help you to take your game to the next level:

Babolat Play Pure Drive racquet

Babolat has been producing tennis racquets for almost 150 years, and has always moved with the times, their products evolving from wooden frames, to metal and then carbon fibres. The influx of smart tech and data analysis in sports is the latest game changer, and Babolat has stayed on the ball here, too, with the introduction of the Play Pure Drive racquet.

Sensors in the handle record every shot that is made, registering the direction of travel and the point of contact, as well as the force, of the ball with the racquet.

Keeping all of the sensors in the handle means that the impact of their weight or positioning on the racquets handle is minimized. A smartphone app acts as a personal coach, analyzing the data collected by the racquet and comparing it with data from other players stored in its database, in order to suggest improvements to your game.  

Sony Smart Tennis Sensor

You don’t have to buy a whole new racquet to benefit from smart tennis technology. Providing you have a compatible racquet, Sony’s Smart Tennis Sensor will simply clip on, allowing you to collect data on every shot. Like the Babolat it comes with its own app which is also a portal to the data collected and collated by other users of the service. Unlike the Babolat, the device can also record video, allowing you to review every shot after the game. This video can be overlaid with graphical visualizations created from the data captured by the racquet, allowing for even deeper insights into a player’s performance.

Adidas MiCoach Smart Ball

This smart football (or soccer ball to Americans) aims to help you improve your play by providing instant feedback on the power and trajectory of your kicks. Like Babolat, Adidas is another old-school sports equipment manufacturer which has consistently moved with the times and clearly sees Big Data and Analytics as the current driving force in sports tech development. The device hides all of its sensors right in the middle of the ball where they won’t affect its dynamics, and transmits them over Bluetooth to its partner smartphone app. It allows free kicks and penalties to be practiced even in a confined area – kick the ball against a wall and the visualizations will show how it would have travelled if you were in the middle of an open pitch.

Zepp Golf

Zepp Labs is a company established with the aim of bringing data analysis into consumer-level sports tech. Their Zepp Golf solution consists of a sensor-enabled glove which is worn during play, and which transmits data on the player’s swing to an analytical smart phone app. One insight which came up early in testing of the product was that older golfers tend to pull back less distance before a swing, resulting in less shot power. The personal coach element of the app monitors a player’s performance for these flaws and suggests remedial action in real-time. The company also produces smart products for baseball and tennis players, and has most recently moved into softball.

Sensoria Smart Sock

A sock might be one of the last gadgets you would expect to see “smartened up” for the Internet of Things age, but you would be wrong!

Sensoria have produced this sensor-stuffed smart sock for runners, which is able to measure not only how far and fast you travel, but the way your foot impacts with the ground – helping you to minimize the risk of stress or injury by maintaining a good form throughout your run. Among its innovations is the sensor technology itself. Textile-based sensors have been developed which can sit between the foot and the running shoe without causing discomfort, and can even be thrown in the washing machine with your regular laundry. Sensoria is another company which was founded specifically to produce sports equipment with analytic functionality, and after raising money for its first product, the sock, via crowdfunding, it has gone on to develop a t-shirt and sports bra also incorporating its textile sensor technology.

Follow us @IoTCtrl | Join our Community

IoT practitioners are at the forefront of their company's digital initiatives. But is the rest of your company ready for its digital moment? The expectations are high in the C-Suite for digital transformations, but there's still more talk than action for many companies.

New research by McKinsey Institute suggests only 17% of corporate boards are participating in strategy for big data or digital initiatives. The good news is almost half of big companies have managed to get their CEOs personally involved, up from 23 percent in 2012.

Other findings from the survey include:

• The most common hurdle to meeting digital priorities, executives say, is insufficient talent or leadership.

• Across the C-Suite, 71% expect that over the next three years, digital trends and initiatives will result in greater top-line revenues for their business, and large shares expect their profitability will grow.

• More than half of executives say that, in response to digital, their companies have adapted products, services, and touchpoints to better address customer needs.

• Executives most often cite analytics and data science as the area where their organizations have the most pressing needs for digital talent, followed by mobile development and user experience.

• Executives who report ample organizational support for adopting risky digital initiatives are twice as likely to work for a high-performing company as executives reporting resistance to risky initiatives due to fear of failure.

• Forty-seven percent say cutting-edge digital work helps them attract and retain digital talent.

• Companies’ priorities vary across industries, reflecting key sources of value in each sector: big data is a top priority in healthcare, for example, while automation is a greater focus in manufacturing (see graphic below).

The digital interconnection of billions of devices is today’s most dynamic business opportunity and at present, the Internet of Things remains a wide-open playing field for enterprises and digital strategy. According to the study, buy-in from the C-Suite and aligning with corporate culture and objectives is key to digital success.

You can read the complete survey here.

Guest blog post by Ajit Jaokar

Introduction

In this series of exploratory blog posts, we explore the relationship between recurrent neural networks (RNNs) and IoT data.  The article is written by Ajit Jaokar, Dr Paul Katsande and Dr Vinay Mehendiratta  as part of the Data Science for Internet of Things practitioners course. Please contact info@futuretext.com for more details

RNNs are already used for Time series analysis. Because IoT problems can often be modelled as a Time series, RNNs could apply to IoT data. In this multi-part blog, we first discuss Time series applications and then discuss how RNNs could apply to Time series applications. Finally, we discuss applicability to IoT.

In this article (Part One), we present the overall thought process behind the use of Recurrent neural networks and Time series applications - especially a type of RNN called Long Short Term Memory networks (LSTMs).

Time series applications

The process of Prediction involves making claims about the state of something in future depending on values in the past and its current state.  Many IoT applications (such as temperature values, smart meter readings etc) have a time dimension. Classical pattern recognition problems are concerned with knowing dependencies between variables (Regression) or in classification of input vectors into categories (Classification).  By including changes in time, we add an additional temporal dimension giving us Spatio-temporal data. Even when a phenomenon is continuous, it can be converted to a Time series by the process of sampling. Thus, phenomenon like speech, ECGs etc can be modelled as a time series when sampled.

Hence, we have a variable x changing in time xt (t=1,2,...) and we would like to predict the value of x at time t+h. Time series forecasting is a problem of function approximation and the forecast is made by computing an error measure over a time series.  Also, given a time series model, we can solve many related problems. For example: Forecast the value of the variable at a time t x(t); Classify the time series at a time in future(will prices go up or down); Model one time series in terms of another(for example Oil prices to Interest rates) etc.

Neural networks for Time series Prediction

However, while many scenarios (such as Weather prediction, foreign exchange fluctuations, energy consumption etc) can be expressed as a time series, formulating and solving the equations is hard in these cases because of reasons such as

a)      There are too many factors influencing the outcome

b)      There are hidden/unknown factors influencing the outcome.  

In many such scenarios, the focus is not to find the precise solution to the problem but rather to find a possible steady state where the system will converge. Neural networks often apply in such scenarios because they are able to learn from examples only and are able to catch hidden and strongly non-linear dependencies.

Neural networks are trained from historical data with the objective that the network will discover hidden dependencies and that it will be able to use them for predicting into future.  Thus, neural networks are not represented by an explicitly given model and can spot nonlinear dependencies in spatiotemporal patterns. They solve the problem of feature engineering i.e. in finding out what is the best representation of the sample data to learn a solution to your problem

Neural networks for time series processing – incorporating the Time domain

When used for Time series forecasting, the obvious first problem is: How to model Time in the  neural network? For traditional applications of Neural networks (such as pattern recognition or classification), we do not need to model the Time dimension. Time is difficult to model in a neural network because it is constantly moving forward.  However, by including a set of delays, we can retain successive values in the time series. Thus, each past value is treated as an additional spatial dimension. This process of converting the time dimension into an infinite-dimensional spatial vector is called embedding. Because for practical purposes, we need a limited set of values – we consider a history of previous d samples (the embedding dimension) as shown in the figure below

Source: http://www.cs.cmu.edu/afs/cs/academic/class/15782-f06/slides/timeseries.pdf.

For an evolution of neural networks, see the previous post Evolution of Deep learning models. Recurrent neural networks are often used for modelling Time series. An example is using Recurrent Neural Networks To Forecasting of Forex(pdf)

A recurrent neural network (RNN) is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behaviour. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. This feature of using internal memory to process arbitrary sequences of inputs makes RNNs applicable to tasks such as unsegmented connected handwriting recognition, where they have achieved the best known results. (Wikipedia). The fundamental feature of a Recurrent Neural Network (RNN) is that the network contains at least one feed-back connection, so the activations can flow round in a loop. That enables the networks to do temporal processing and learn sequences, e.g., perform sequence recognition/reproduction or temporal association/prediction. Using the same idea as time delays as above, the recurrent neural network can be converted into a traditional feed forward neural network by unfolding over time as shown below.

Source:  http://www.cs.bham.ac.uk/~jxb/INC/l12.pdf

RNNs are used to model Time series because the feedback mechanism creates a ‘memory’ i.e. an ability to process the Time dimension. Memory is important because many Time series problems (such as Traffic modelling) need a long term / historical modelling of Time values.  Long Short Term Memory networks(LSTMs) are a special kind of RNN, capable of learning long-term dependencies especially because they are capable of remembering information over long time frames. The figure below summarises feed forward neural networks and Recurrent neural networks.

Source: http://deeplearning.cs.cmu.edu/notes/shaoweiwang.pdf

Implications for IoT datasets

In subsequent articles, we will explore LSTMs in greater detail and implications for IoT data.

This article covers topics we teach in the Data Science for Internet of Things practitioners course. Please contact info@futuretext.com for more details

 Follow us @IoTCtrl | Join our Community

The smart phone on your belt is dramatically different from the flip phones of a decade ago. Technology continues to move at incredible speeds and we are truly living in a golden age. But the where we are headed is unlike where we’ve been.

In the future, the Internet of Things will be a reality in every sector. Smart systems will be released with sensors and robotics that simplify and automate manufacturing. The system will operate through wired and wireless networks and an infrastructure will help us to accomplish more during the course of a day.

This begins with physical objects, built with sensors and actuators placed in them. These individual parts will send and receive information in order to complete specific tasks. They will depend on real time data and this information will affect the big picture. In fact, each device on the assembly line will connect to a central system that will orchestrate and synchronize the entire system to ensure things run smoothly and as effectively as possible.

In order for smart manufacturing to work, there need to be systems in place that work with the smart manufacturing vision. Sensors must be placed in technology and a host system installed. This will help with logistics, order placement, procurement and other essential functions that impact the overall system.

So who does this? While your IT department could technically handle the task, it would be time consuming and cost you hundreds of man hours to develop. A better choice is to consider a vendor who can help with the effort. These individuals will help to create a functional system which is tightly integrated and allows you to effectively manage your manufacturing operations. With new industry standards being released for manufacturing all the time, it is certain the internet of things will play a pivotal role in the future of manufacturing automation.

An example of it is already seen in the food and beverage industry. Machines currently communicate sensitive information like temperature, humidity and the condition of the containers. Companies can also track shipments with identifying codes and determine where they originated from in the company and where these items were shipped to in the world. If there is a case of contamination, they can also quickly contact locations who received items that might be tainted.

When the internet of things becomes dominate on these manufacturing lines, there will be more power. There will be a central master computer that will run the entire operation. It will have an intelligent way to analyze, address concerns and to remain independent at all times, all while continuing to meet the demands of production.

There is no denying the internet of things will play an important role in the future of production. Good will be released faster and profits will spike for a company. That makes it important to embrace today and incorporate in the current structure of your business. Doing that will help you to be part of the future and to remain a visionary in the industry.

Are you hiring ahead of the coming shift in how workers work?

How Did We Create Such a Rich Market?

Want to know how the "Internet of Things" became a thing at all? To do so, you must look back to the start: the birth of networking and the explosion of consumer technology.

The internet isn’t that old, so far as the world wide web. In 1974, the structure we know and love today was born. Just ten years later. that the first domain name system was introduced, allowing for easier networking. The first website actually came online in 1991. The "internet," as a network of connected devices in consumer homes, was only proposed just a scant two years before that, yet it came crashing into our mainstream world. 

In no time the internet took over. By 1995, multiple websites and systems came online. I remember watching crude bulletin board systems arise, then quickly be replaced by Geocities pages and early websites. The first business webpages actually came in the form of reproduced fliers, essentially scanned and put online to promote companies. All of these new ideas came from the imaginings of others that had taken place decades earlier.

The term “internet of things” or “IoT” is also not a new one. You can find references to it as far back as the idea of the Internet itself, but if you survey an IoT team, it is more than likely that few know this. The history, or at least the ideology, goes back a great deal further than most people know. This, of course, has ramifications on the marketplace, both in how older technology companies approach the space and how traditional product introduction processes operate.

Thinkers across history could be responsible for coining the term, depending on the story you read. Some point to Tesla and Edison as the first to lead connected objects. Others look at the literal applications by Tim Berners Lee and Mark Weiser, the latter of which famously created a water fountain synced to the activities of the NYSE. The founders of Nest could also make the list, one of the first truly non-computer connected objects.

Even the idealism and futurism of the 1950s and 1960s gave way to the Internet of Things thinking. Imagine a classic 60s technology ad, displaying the "home of the future." Everything is connected and communicating, and people are never out of reach of their day-to-day technology.

Then, of course, is Kevin Ashton, a man who comes up when you Google "who came up with the Internet of Things." Kevin is a frequent thinker in the space who is corrected attributed to a verifiable creation of the term, "Internet of Things." Like most corporate lingo, the origin is likely impossible to pin down, but the idea that the term was born in a boardroom is not surprising. The leaders who would go on to actually take these objects to market in the 90s included "traditional" players like IBM and Sony.

The story is that, no matter what route you pick to decipher the past, the rise of Internet of Things thinking is ubiquitous. From the moment "networking" arrived into everyday life, people were thinking about how it would impact our world.

1998 itself is a turning point in many ways, when something changed. Apple returned to the market with the iMac, and the team that designed this platform would go on to design the iPhone and, most critical to IoT research, the iPod. Big name manufacturers that had for most of their development focused on the PC were now investing in everyday objects with connectivity and technological features. The smartphone era was planted, and with it would come the first real consumer-level IoT object based on existing computers.

The history of IoT is extraordinarily dense, and the reading of the history depends on who you ask. If you were to question a designer at IBM in the late 1980s, you would find ideas similar to what we now call IoT in constant use. However, if you ask an emerging startup from the early 2000s, you would find a wave of thinkers taking credit for the idea. The reality is somewhere in between: those who thought ahead about computers expected what we have today, billions of devices.

IoT has continued to grow and to evolve and projections are bright for this new methodology for using the internet. The future of IoT is now –with devices coming online every day. The world is reliant upon connected cars, connected medical devices and even connected homes.

Companies today are scrambling to get their own IoT systems online and moving, and new recruits are being brought in every day to head up IoT systems in companies from small to large. How well do they know the history of the space and exactly how broad it can be?

 We want your input - please share your thoughts below!   Click Here 

By Venkat Viswanathan and Ravi Ravishankar

As the Internet of Things (IoT) gains momentum, it’s apparent that it will force change in nearly every industry, much like the Internet did. The trend will also cause a fundamental shift in consumer behavior and expectations, as did the Internet. And just like the Internet, the IoT is going to put a lot of companies out of business.

Despite these similarities, however, the IoT is really nothing like the Internet. It’s far more complex and challenging.

Lack of Standardization

Unlike the Internet, where the increased need for speed and memory was addressed as a by-product of the devices themselves, the sensors and devices connecting to the IoT network have, for the most part, inadequate processing or memory. Furthermore, no standard exists for communication and interoperability between these millions of devices. Samsung, Intel, Dell and other hardware manufacturers have set up a consortium to address this issue. Another equally powerful consortium formed by Haier, Panasonic, Qualcomm and others aims to do the exact same thing. This has raised concerns that each of these groups will engage in a battle to push their standard, resulting in no single solution.

New Communication Frontier

The Internet was designed for machine to human interactions. The IoT, on the other hand, is intended for machine-to-machine communications, which is very different in nature. The network must be able to support diverse equipment and sensors that are trying to connect simultaneously, and also manage the flow of large quantities of incredibly diverse data...all at very low costs. To meet these requirements, a completely new ecosystem—independent of the Internet—must evolve.

Data Privacy

The IoT also raises serious challenges for data security and privacy. Justified consumer concerns will call for stricter privacy standards and demand a greater role in determining what data they will share. These aren’t the only security issues likely to arise. In order for a complete IoT ecosystem to emerge, multiple players must use data from connected devices—but who owns the data? Is it the initial device that emits it, or the service provider that transports that information, or the company that uses it to provide the consumer better service offerings?

Geographic Challenges

For multinational organizations with data coming from various regions around the globe, things get even more complicated. Different countries have different data privacy laws. China and many parts of the EU, for example, will not let companies take data about their citizens out of their borders. This will result in the emergence of data lakes. To enable business decisions, companies must be able to access data within various geographies, run their analysis locally and disseminate the insights back to their headquarters…all in real-time and at low costs.   

In spite of all these challenges, the IoT is not something companies can afford to keep at arm’s length. Like the Internet, it will empower consumers with more data and insights than ever before, and they in turn will force companies to change the way they do business. From an analytics perspective, it’s very exciting. Companies will now have access to quality data that, if they combine it with other sources of information, can provide them with immense opportunities to stay relevant.

As an example, let’s look at the medical equipment industry. Typically these companies determine what equipment to sell based on parameters like number of beds and whether the facility is in a developing or developed market. However, these and other metrics are a poor substitute for evaluating need based on actual use. A small hospital in a developing country, for example, will diagnose and treat a much wider range of diseases than a similar facility in a more developed region. By equipping the machines with sensors, these manufacturers can obtain a better understanding of what is occurring within each facility and optimize selling decisions more effectively as a result.

This is just one example to underscore the tremendous potential that the IoT holds for businesses. In order to truly realize these and other opportunities, companies must understand the challenges outlined above and have a framework in place to address them. In the early days of the Internet, few could have predicted its transformative impact on all facets of our lives—personal and professional. As the IoT heads into its next phase of maturity, we can expect to see a similar effect emerge.

Ravi Ravishankar is Global Head of Product Marketing and Management at Equinix's Products, Services and Solutions Group and Venkat Viswanathan is Chairman at LatentView Analytics.

Originally Posted on Data Science Central

Follow us @IoTCtrl | Join our Community

The Internet of Things encompasses a wide range of connected services, technologies, and hardware devices. Yet, for consumers, it is the growing number of portable and wearable devices that will be their main interface with IOT tech. The wearable device market is rapidly evolving, especially when it comes to smart watches and fitness monitoring devices.

As opportunities grow, the wearables dominating the market are also changing. What does this mean for those involved in the development, marketing, and sales of these IOT connected devices?

 How Big is the Wearable Market in 2015?

International Data Corporation (IDC) has predicted that wearable device shipments in 2015 will rise to 173% of the total sales achieved in the previous financial year. This translates to over 72 million devices, including smartwatches and health trackers. This growth has been largely driven by high profile releases such as the Apple Watch in April of 2015, and also by widely publicized financial opportunities, Fitbit’s recent IPO being a prime example.

With the potential to move over 72 million units across the market, it is no surprise that leading technology companies like LG, Samsung, Sony, Microsoft, Apple, and Motorola are starting to increase their focus on wearable technology.

When we look closer at the marketplace, we see a strong mix of upstart companies and traditional players, with Fitbit, Garmin, and Xiaomi all new entrants. This blend of "old" technology giants and very new companies is promising - the marketplace is growing rapidly, and opportunity actually exists.

Future growth will be an incentive for further investment. IDC figures suggest that by 2019, global sales of wearables could exceed 150 million units. The market is open completely, with any company able to take a device to market open to growth.

Do these figures mean success for all involved in the wearable market? Not entirely.

Challenges for Businesses to Adapt

Although the overall market has grown, recent trends show that wearable fitness devices are losing out to increased smartwatch sales. Gartner’s latest research suggests that the dip could largely be associated with the increasing crossover in functionality between fitness devices and the latest smartwatches. 50 percent of those seeking a fitness wearable will end up choosing a smartwatch instead, and brands do not necessarily know why this shift is happening.

I think that one feature overlap is contributing to this. Fitness devices chiefly collect information relating to distance covered, physical location, and heath, including heart rate. Nearly every smartwatch on the market today can do all of this, and more. For a savvy consumer, combining a Samsung Galaxy Gear smartwatch with a high-end Galaxy Note 4 or Galaxy S6 would provide GPS tracking, information on calories burnt, heart rate monitoring, and even blood oxygen levels. The technology is advancing year on year, and it is clear that the innovation gap is already closing.

There are two consequences I see with this lack of clear differentiation. The first is that fitness-focused products need to innovate or die. With the market contracting by supporting multi-feature devices over purpose-built tools, the new goal should be for innovation to differentiate. Put simply, the fitness trackers of the world need to do something that smartwatches cannot.

The second consequence is that companies like Fitbit and Nike, which are focused on fitness tracking, will need to lower prices to compete with integrated smartwatches. When a consumer is faced with a $120 fitness tracker and a $200 smartwatch with phone connectivity, alerts, and apps, the choice becomes very one-sided. Yet, the bottom of the market, and the sector more likely to actually increase sales of purpose-built trackers, is relatively unsaturated. 

Fitbit, Jawbone, and Nike make up 97% of the wearable fitness device market. In smartwatch territory, it is Samsung and Apple that lead the market. Looking at one of the least expensive fitness trackers, Fitbit's Zip, we see a $60 base price point. Even at this level, the casual user has to pause and think - their phone already does much or all of what the Zip does, and a waterproof fitness case is cheaper. Fitbit, in this case, needs either to more fundamentally differentiate or drop its pricepoint.

Where is the Money in Wearables?

Even with staggering sales numbers, wearables are not in themselves a key revenue stream. Instead, it is the associated value that provides the biggest benefit to manufacturers.

Smartwatches, in particular, are seen as accessories. They are paired to smartphones and in turn can help to drive sales. They are also showpiece items. Even if Samsung, Apple, Sony etc. only manage to sell wearable technology to 10% of their smartphone customers (a speculative number), they will generate brand marketability, and logically would experience knock-on sales.

When it comes to companies like Nike, Fitbit, and Jawbone, the profit can come from connected services. Examples include subscription based exercise plans, analytics software, and in the case of Nike, a wearable can lead to increased apparel sales.

Still, there is an incredible gap for new entrants to the market. Apple and Samsung can rely on a massive pool of existing customers, and directly integrate their offerings into that group. Fitbit cannot, with no "hub" devices on the market. Even subscription-based models cannot make up for the gap. This makes the marketplace incredibly hard to predict going forward - nothing prevents a company like Samsung from releasing another mid-range watch and completely dividing the market. 

As with all IOT technology, the wearable device is only one part of the experience, and therefore only one part of the business model. It is the way in which data is collected, analyzed, and presented that provides the true value of any smart device. Smartwatches already have an advantage because they are highly integrated into their respective smartphone operating systems. Wearable fitness device companies have the opportunity to provide fitness tracking as a service, and must find new ways to monetize the service to generate direct revenue on top of initial hardware sales.

What does the Future Hold For Wearable Technology?

Over a billion smartphones were sold around the world in 2014. Global wearable sales make up less than 10% of that number. The challenge for manufacturers is to develop wearables that easily integrate with daily life that also are something that consumers want to use on a daily basis.

While wearables are high in consumer mindshare, they are relatively low in actual penetration. Smartwatches are now able to integrate a fitness device with a smart device in a way that is both compelling and practical, but is it enough? Those in the industry will need the best ideas, the best strategies, and the best talent to ensure that in-demand products are developed in line with business goals, and that they result in strong financial growth.

When considering how to hire leadership for the emerging Internet of Things market, keeping these consideration in mind is critical.I can help guide your choices, find the best candidates, and bring IoT experience to your company. Contact me today for a consultation.