ATHARVA : Autoregressive Trend analysis & Hypercritical Air pollution Ranging & Visualisation Application

Aim : To develop AI/ML-based software to identify/analyze: 1. Location of hot spots. 2. The long-term occurrence of hot spots and changes.

We have developed ATHARVA which is a user friendly Web application that can be used by common people to gain the knowledge about hotspots of India and their changes.

Link to PPT:

https://www.canva.com/design/DAEDV70XfpA/Qmn-R7Y6_iSq63JsldarnQ/view?utm_content=DAEDV70XfpA&utm_campaign=designshare&utm_medium=link&utm_source=sharebutton#1

Idea Short Video:

https://drive.google.com/file/d/1jaEzWIA169S9Ln3C83RpgdI0CYdDdkm3/view?usp=sharing

Tweet Link:

https://twitter.com/darecoder/status/1285983011855454208

Features in our application Atharva:

• District-wise plotting Heatmap for each pollutant analyzation and visualization
• Coordinate basis Hotspots and Coldspots detection
• Graphical Analysis of concentration of pollutants at each hotspot
• Filters available for querying various pollutants on the map for any week.
• Plotting shifts in HotSpots over a specific period of time
• Analyzing those shifts by plotting backward trajectories and verifying it using meteorological data.
• Two modes for visualization: Dark mode & Daylight

Steps to setup project(ATHARVA) locally

For Linux OS:

1. NPM install

npm install
npm run clientinstall

2. Install Postgres & Postgis using apt

• Create new user and database
• Switch to postgresql cli mode then, grant superuser privileges to your newly created user. (ALTER USER your_username WITH SUPERUSER;)
• Start postgres by running: systemctl start postgresql
• Replace user, password, database with the user, password, database (name) created above in knexfile.js.

3. Run knex migrate:latest

knex seed:run
npm run dev

Updates on project till now

• Fetched the sentinel-5p data.
• Performed data mining and data preprocessing.
• Plotted Hotspots district-wise using spatial auto-correlation.
• Analysed the movement of Hotspots.
• Plotted backward trajectory by using metrelogical dataset.
• Depicted the shifts in hotspot over a period of time.
• Trend analysis at each hotspot for concentration of every pollutant gas.
• Multiple Heatmap plotted using MapBox for better visualization.

Tools and Technologies used

• Sentinel-5P Pre-Operations Data Hub (data source)
• HARP (for fetching Sentinel5p data)
• PYSAL (for spatial relationship to detect hotspots, coldspots)
• ESDA (for exploratory analysis of spatial data)
• GEOPANDAAS (to work with geospatial data and geojson data)
• METEX (for determining trajectories through backtracking analysis of metreological data)
• NCEP/NCAR (data source for metreological data needed for trajectories)
• React,Node.js,Express.js,Postgis,PostgreSQL (for our Web Application)

ATHARVA UI:

YouTube Video:

January, 2020 Clusters & Cluster centers:

February, 2020 Clusters & Cluster centers:

Change in HotSpots & Moving HotSpots:

HotSpot Trajectories: