Subscribe to DSC Newsletter

Data Science for Internet of Things - The Big Picture

This big picture view lays the foundation of our book Data Science for the Internet of Things. (Co-authored by Ajit Jaokar, Jean Jacques Bernard and Sukanya Mandal)

We address the question: at what points can we add analytics to the data after it leaves the sensor and what are the implications of doing so at various stages.

In this diagram, we present the big picture through two process flows:

  1. Technology flow: Edge to Stream to Store
  2. Deployment flow: Model build, deploy and refresh in production including at the edge

Data Science for IoT implementation differs from traditional Data Science in four key aspects

  1. Edge Computing
  2. Feature Engineering for IoT
  3. Complex event processing
  4. Embedded AI

The last three are not shown in the diagram to make it more readable. 

 

Notes:

  1. From an IoT analytics perspective, Mobile devices could also function as Edge devices.
  2. AI when deployed at the edge works primarily for inference(currently)
  3. Most Data Science for IoT problems are time series problems. In practise, this means the use of LSTMs in many cases. We could also work with images, sound etc (using convolutional neural networks etc)
  4. The diagram does not show batch mode of processing which exists in many applications

In terms of deployment models, we consider three options:

  1. Digital signals: for example, in insurance where IoT provides a signal to enhance existing business processes
  2. Digital twin: Mainly in industrial iot
  3. Digital wand: A ‘pervasive’ deployment of AI / ubiquitous computing. Currently, seen in some smart city applications in China

We welcome comments

To sign up for the book chapters and ongoing emails https://my.sendinblue.com/users/subscribe/js_id/31hme/id/1%20future...

To learn about my course on AI please see #AI/ #Deep Learning applications course

Views: 13194

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by ajit jaokar on May 28, 2018 at 4:02pm

@jaap thanks for this. Its good. I see the two streams (monitoring and Provisioning) and then the CRISP-DM around it. Good idea :)  for my diagram, I struggled to add details and keep it readable, rgds ajit

Comment by jaap Karman on May 27, 2018 at 9:06am

Ajit, sorry  there is no URL as this picture is not published anywhere yet. It is a part of a PPTx diagram no 9 (out of 19) as of  my brain dump how to do a complete AI proces with alignment on all kind of levels.
Starting wiht the common seen problems wiht modelling en blind handover to IT.

I started wiht the well known Crisp_dm circle adding all kind of compliancy and process requirements and ended with this:

The handing over of the three green lines … gettin raw data tranform to  usuable data score-publish is the connection form AI to generating value als run by ICT. 

Comment by ajit jaokar on May 27, 2018 at 8:31am

@jaap thanks. very useful diagram. whats the URL of that diagram ie source(I clicked on it also) 

agree that two dedicated lines of development will become standard

Comment by Vincent Granville on May 27, 2018 at 8:21am

To get a higher resolution of the picture in Jaap's comment, click on it.

Comment by jaap Karman on May 27, 2018 at 8:13am

Nice post but I don't like the Docker there as it seems to be an blind handover to It guys.
The thinking in two dedicated lines for developement and aplorarations shoudl become a standard.
The same kind of thinking…. (no no book promotion just experiences) 
Note the evaluation of mode behavior as it is creating new data and is data on its own. 

Follow Us

Videos

  • Add Videos
  • View All

Resources

© 2018   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service