How nishantrajora/aqi-prediction-india-ml works
This repository presents an AQI forecasting project: historical city-level observations are explored and preprocessed, transformed into time-ordered training and test data, and used to evaluate Random Forest and XGBoost models. Its Streamlit app communicates the project’s EDA, preprocessing, scores, and feature-importance findings. The README describes a deployed prediction dashboard, but the inspected app source renders explanatory content and metrics; no live inference path is evidenced. Notebook internals were not sampled, so their detailed responsibilities and connections are based on the README.
Main entry points: preprocessed_city_day.csv Prepared records · model_and_testing.ipynb Model evaluation · web_app.py Streamlit dashboard · city_day.csv City AQI records
Components 12
AQI Data
- City AQI records
main_project/city_day.csv - Prepared records
main_project/preprocessed_city_day.csv - Training features
main_project/X_train.csv - Training targets
main_project/y_train.csv - Test features
main_project/X_test.csv - Test targets
main_project/y_test.csv
Analysis & Modeling
- EDA workbook
main_project/EDA_Tableau.twb - Preprocessing
main_project/Preprocessing.ipynb - Model evaluation
main_project/model_and_testing.ipynb
Project Dashboard
- Streamlit dashboard
main_project/web_app.py - Histogram collage
main_project/collage_making.py
Other components
- Project visitor
Connections 14
- City AQI records → EDA workbook: explores trends
README.md - City AQI records → Preprocessing: preprocesses
README.md - Preprocessing → Prepared records: prepares records
README.md - Prepared records → Training features: time-splits
README.md - Prepared records → Training targets: time-splits
README.md - Prepared records → Test features: time-splits
README.md - Prepared records → Test targets: time-splits
README.md - Training features → Model evaluation: trains models
README.md - Training targets → Model evaluation: trains models
README.md - Test features → Model evaluation: evaluates models
README.md - Test targets → Model evaluation: evaluates models
README.md - Project visitor → Streamlit dashboard: views dashboard
main_project/web_app.py - Model evaluation → Streamlit dashboard: shows reported results
README.md - Histogram collage → Streamlit dashboard: supplies histogram display
main_project/web_app.py