Lots Collected, Little Analyzed

  • Collection without analysis has negative ROI

 

Most IoT data are not used currently. For example, only 1 percent of data from an oil rig with 30,000 sensors is examined.

-- McKinsey & Company

 

  • As difficult as it is to collect and capture data,
    gaining insights via analysis takes just as much effort

 

  • Dependable, real-world on-line analytics take time to develop, tune, and test.

A Project Focused Tour

Tide Prediction - Online Analytics

Nighttime Noise - Historical Analytics

Weather - IoT Analytic Workflow

Counting Cars - Edge Node Analytics

A Project Focused Tour

Tide Prediction - Online Analytics

Nighttime Noise - Historical Analytics

Weather - IoT Analytic Workflow

Counting Cars - Edge Node Analytics

Weather - IoT Analytic Workflow

MathWorks Weather Station -- Revisited

Classic IoT Maker Project with Machine Learning

Deep Dive: Google "mathworks weather revisit"

http://goo.gl/TVqiOD

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

Weather - IoT Analytic Workflow

MathWorks Products for IoT

MathWorks Products for IoT

ThingSpeak.com

  • Analytic IoT Platform                      
  • Running for more than 6 years    
  • >20,000 active users worldwide  
  • Integrated MATLAB engine            
  • Designed for engineers and scientists                                         
  • Free for low data volumes                    
  • Lots of examples and projects

A Project Focused Tour

Tide Prediction - Online Analytics

Nighttime Noise - Historical Analytics

Weather - IoT Analytic Workflow

Counting Cars - Edge Node Analytics

Car Counter - Edge Node Analytics

Counting Cars

Traffic monitor using a Raspberry Pi & Webcam

Deep Dive: Google "count cars thingspeak"

http://goo.gl/KGdofK

Car Counter - Edge Node Analytics

Counting Cars

Traffic monitor using a Raspberry Pi & Webcam

Deep Dive: Google "count cars thingspeak"

http://goo.gl/KGdofK

Car Counter - Edge Node Analytics

Embedded devices have always used:

  • Signal processing
  • Matrix math
  • Closed loop control

 

Powerful mobile processors enable:

  • Sensor fusion & analytics
  • Computer vision & image processing
  • Trained machine learning
  • Statistical modeling
  • Small vocabulary voice recognition

Car Counter - Edge Node Analytics

Car Counter - Edge Node Analytics

Foreground Detection, 2-D Median Filter, Blob Analysis

Feeds custom counting block and Transmit to ThingSpeak

Car Counter - Edge Node Analytics

Foreground Detection, 2-D Median Filter, Blob Analysis

Feeds custom counting block and Transmit to ThingSpeak

Car Counter - Edge Node Analytics

A Project Focused Tour

Tide Prediction - Online Analytics

Nighttime Noise - Historical Analytics

Weather - IoT Analytic Workflow

Counting Cars - Edge Node Analytics

Night Noise - Exploratory Analysis

Night Noise Analysis

Identify unusual changes in ambient noise level at night

Night Noise - Exploratory Analysis

Median Filter

Night Noise - Exploratory Analysis

Normalized based on first 3 hours
Polynomial fit of expected noise levels

Night Noise - Exploratory Analysis

Identify alarm conditions

> 1 std deviation above expected

Night Noise - Exploratory Analysis

To do: Investigate what happened this night...

 

Night Noise - Exploratory Analysis

A Project Focused Tour

Tide Prediction - Online Analytics

Nighttime Noise - Historical Analytics

Weather - IoT Analytic Workflow

Counting Cars - Edge Node Analytics

Tide Prediction - Online Analytics

On Line Tide Alerts

Tide Measurement and Prediction

Deep Dive: Google "tide matlab thingspeak"

http://goo.gl/uXHkYU​

Tide Prediction - Online Analytics

Edge Node

  • Arduino                                            
  • Ultrasonic Rangefinder and Temperature Sensor                          
  • Data Reduction                                   
  • De-noising and "De-batting"          
  • Sends Data to ThingSpeak

Tide Prediction - Online Analytics

On-Line Analysis

  1. Median Filter
  2. Temperature Compensation
  3. Convert to Water Depth
  4. Tweet @tidealerts
  5. Hysteresis and state machine to avoid being too "tweekative"

Tide Prediction - Online Analytics

On-Line Analysis

  1. Median Filter
  2. Temperature Compensation
  3. Convert to Water Depth
  4. Tweet @tidealerts
  5. Hysteresis and state machine to avoid being too "tweekative"

Tide Prediction - Online Analytics

Thank You!

  • Collection without analysis has negative ROI
     
  • Analysis happens at every stage
     
  • Dependable, real-world on-line analytics take time to develop, tune, and test
     
  • Trust can be lost in a single moment                                      
  • Check out ThingSpeak.com!

http://slides.com/rpurser/environmental_analysis

Made with Slides.com