AirVision
AirVision: AQI Prediction & Analytics Platform
Eleven regressors, one benchmark, a live AQI dashboard.

Abstract preview - not a screenshot
01 / Problem
Predict the Air Quality Index from pollutant measurements across multiple Indian cities.
02 / Approach
- 01Imputation, date features, encoding, scaling, evaluation and saved artifacts.
- 02Benchmarked 11 regression algorithms: Random Forest, XGBoost, LightGBM, CatBoost, SVR, KNN, AdaBoost, ElasticNet, Decision Tree, Gradient Boosting, Linear Regression.
- 03Interactive Streamlit dashboard.
03 / Architecture
04 / Key details
- AQI prediction from pollutant measurements across multiple Indian cities.
- Benchmarked 11 regression algorithms.
- Best: CatBoost, R² = 0.91, MAE = 12, RMSE = 19.
- Interactive Streamlit dashboard.
05 / Results & signals
- R² 0.91
- CatBoost
- 11
- Models benchmarked
- MAE 12
- Mean abs. error
- RMSE 19
- Root mean sq. error
06 / Links
That's every entry · back to the index
Back to work
Return to the home page, right where you left the project grid.