This repository contains a project by the "Why are you painting that fence again?" team we created during the Raifhack DS hackathon. The hackathon's goal was to build a model for predicting the value of commercial real estate. We placed 28th out of nearly 400 teams.
To run the model, you need to do the following:
- Go to Eden directory.
cd eden
- Set up environment using poetry
poetry install
- Load data from W&B
poetry run python3 ../read_data.py
- Run preprocessing
poetry run python3 preprocess.py -i data/split_data -o data/some_preprocessed_data
- Train the model
poetry run python3 train.py -t0 <t0 path> -t1 <t1 path> -mp saved_models/baseline_changed.pkl
<t0 path> -- path to csv file with price_type = 0.
<t1 path> -- path to csv file with price_type = 1.
- Predict
poetry run python3 predict.py -p1 <test path> -mp saved_models/baseline_changed.pkl -o submissions/baseline_changed.csv