GeoAI-Challenge-for-Air-Pollution-Susceptibility-Mapping

Submissions for username JuliusFx for the above ITU problem statement

This repository contains the following files

1. Technical Report - gives the summary of the model developed

2. main.ipynb- the jupyter notebook used to generate the predictions

3. The datasets provided

   a) Train.csv-contain train features (this file is big cannot upload here-- download from data section of page url provided below)

   b) Test.csv-we are supposed to predict the aqi values for the 160 location in the city of milan

   c) air_pollution_Milan_Comune_topo_only.csv-contains topological features for the city of milan

   d) Interpolated seasonal datasets from the city of milan from a regular grid

     i) meteo_autumn_2022
   
     ii) meteo_spring_2022
   
     iii) meteo_summer_2022
   
     iv) meteo_winter_2022
   

4. requirements.txt- has the module version required to run the notebook successfuly

5. lgb_sub.csv - contains the predictions for our proposed model

6. Grand Final Presentation Power point - GeoAI Air Quality Susepcitibility.pptx

Order of Running the Files

1. Ensure all files are residing in the same folder. I run them on conda environment in my computer

2. Install the models in requirements.txt

3. Run main.ipynb-should be able to generate the lgb_sub.csv file

To view the competition page where you can get more information on the problem statement, please visit: https://zindi.africa/competitions/geoai-challenge-for-air-pollution-susceptibility-mapping