1. QUICK START: Run TWO TML Solutions Right Now!
Important
The power of TML is not only in how it can process and perform machine learning at the entity level, in-memory, but the amazing real-time visualizations that users can create with the TML output for faster, deeper, insights from real-time data streams.
1.1. QUICK START: TML Solution with Real-Time Entity Based Processing
For users who want to quickly see a running solution now, just do the following.
Note
You must have docker installed.
1.1.1. Run this docker command
docker run -d -p 9005:9005 maadsdocker/seneca-iot-tml-kafka-amd64
Tip
Wait about 10 seconds…
1.1.2. View The Real-Time Dashboard
Then, open up your favorite browser and enter this URL below:
http://localhost:9005/iot-failure-seneca.html?topic=iot-preprocess2,iot-preprocess&offset=-1&groupid=&rollbackoffset=500&topictype=prediction&append=0&secure=1
Tip
PRESS THE RED “START STREAMING” button in the top-left…
You should see this Dashboard in your browser start to populate with real-time preprocessed IOT data:
Note
The above dashboard is processing real-time data and streaming it directly from your container to your browser using websockets.
Tip
Hover over with your mouse on the map bubbles. You can also download all the table data by clicking “Download as CSV”.
1.2. QUICK START: Another TML Soluton with Real-Time Entity Based Processing AND Machine Learning
Let’s run another TML solution, but this time with machine learning models being created for each device entity.
1.2.1. Run this docker command
docker run -d -p 9006:9006 maadsdocker/uoft-iot-tml-kafka-amd64
Tip
Wait about 10 seconds…
1.2.2. View Another Real-Time Dashboard
Then, open up your favorite browser and enter this URL below:
https://localhost:9006/iot-failure-machinelearning-uoft.html?topic=iot-preprocess,iot-ml-prediction-results-output&offset=-1&groupid=&rollbackoffset=500&topictype=prediction&append=0&secure=1
Tip
PRESS THE RED “START STREAMING” button in the top-left…
You should see this Dashboard in your browser start to populate with real-time entity based probability predictions of IOT device failures. The figure below shows 43 machine learning models created for 43 devices!
Tip
Press the TOGGLE button in the top-right of the dashboard.