The IoT Show


Wednesday, November 04, 2015

Data Science and The Interent of Things

Image Credit: TechTalk on YouTube

I love the junction of Product and Data. Internet of Things creates value at that perfect junction making it exciting for me.

IoT is about making ordinary things all around us smart by adding sensors and giving it Internet access. These sensors keep collecting data. It could be the temperature inside your home, oven, our bodies, or in a steel plant or many other sensor data for motion, luminescence, touch, acceleration, humidity etc. The sensors collect this data at a repeat frequency creating a vast pool of data.
In simple cases the sensor can check against a preset level and take some action. Turn on the sprinkler if the temperature is above a certain threshold. Notify the granary owner if the humidity of stored wheat is above a certain safety level. Send me an alert if my garage is left open for longer than five minutes.

As I play with more IoT products and engage with entrepreneurs solving customer problems I am increasingly fascinated by the power from processing IoT Data and the many ways value is created from this data.

Data Science and IoT from Ajit Jaokar
Ajit Jaokar was the first person who introduced me to the power of Data Science and IoT. He has taught at Oxford, is based in London and offers online classes customized to anyone who wants to learn this topic either to solve a complex data problem for their business or to enhance their careers in IoT Analytics.

In simple cases, one can look at IoT data and analyze for patterns to build predictive models. This is used by Boeing and many manufacturing plants to monitor health of equipment to predict failures to do preventive maintenance.

Machine Learning
Instead of rule setting you could apply Machine learning and now the horizons open up.

Machine Learning involves training a model by feeding it data to make predictions against a performance measure.

Nailbot is a robot that offers manicures for teen girls and uses machine learning to adapt to nail sizes of the girls.  ProGlove, an industrial wearable uses machine learning to predict any quality glitches in a manufacturing process.

Take it another notch with Deep Learning
Deep learning is about feeding large volume of data to the computer to learn about a system or game or character identification without a human telling them the rules. It is best suited for problems that have many dimensions and no rules. e.g Facial recognition
Ajit is passionate about the idea of applying deep learning to smart city scenarios. Imagine being in a crowded stadium or 4th of July parade and if a parent gets separated from a child, its a nightmare to locate the child. Deep learning of the video footage can apply the algorithm similar to the one Apple or Google or Facebook use to identify faces in photos to predict where the parent and child are in the crowd in real time to help bring them together.

Ajit is offering the  3rd cohort of his Deep learning and IoT class onsite for Londoners or Online for everyone else (across time zones) and the next batch starts on Nov 10th - Details here.
I would love to hear your comments below if you have taken Ajit's Data Science and IoT course or if you have questions/comments/skepticism about how real is IoT and Data Science or if you are using Data Science for your IoT product to do something fascinating.
Sudha Jamthe is the author of the Amazon Kindle book "IOT Disruptions" anda Mobile & IoT Product and Business Leader who loves playing with IoT and data to figure out whether we can build products to change the boundary of human and machines. She is an innovation advisor for Blockchain University and Barcelona Technology School. Join her weekly video show The IoT Show for analysis with global guests about innovations in Internet of Things.
Opinions are mine unless you want to own them

1 comment:

Standard Solutions said...

nice and interesting page..thanks for posting such a great information.. keep posting such articles...
SEO Expert