A Spatio-Temporal Framework for Real Estate Appraisal

This repository is an official PyTorch implementation of the paper

STRAP: A Spatio-Temporal Framework for Real Estate Apprisal., Lee et al., CIKM 2023 (Short).

Requirements

Download dataset

Download the real estate dataset from the below link and unzip below /data/preprocessed/

real estate data (~100mb)

PWD: 1234

Install Anaconda environment

We assume you have a Linux device to run these scripts. First, install an anaconda environment to train the deep learning model.

wget https://repo.anaconda.com/archive/Anaconda3-2020.07-Linux-x86_64.sh
bash Anaconda3-2020.07-Linux-x86_64.sh
source ~/.bashrc

Install dependencies

We assume you have access to a GPU that can run CUDA 11.1 and CUDNN 8. Then, the simplest way to install all required dependencies is to create an anaconda environment by running

conda env create -f requirements.yaml

After the installation ends you can activate your environment with

conda activate strap

Then further install few more dependencies

pip install hydra-core --upgrade
pip install opencv-python

Signup Wandb

https://wandb.ai/home

Instructions

Check data preprocessing

If you want to see the data pre-processing procedure, check the below jupyter notebook scripts.

./data/dataset_summary.ipynb

Train model

If you want to train the model on your own, use the main.py script

python main.py

Citation

@article{lee2023strap,
  title={ST-RAP: A Spatio-Temporal Framework for Real Estate Apprisal},
  author={Hojoon Lee and Hawon Jeong and Byungkun Lee and Kyungyup Daniel Lee and Jaegul Choo},
  journal={CIKM},
  year={2023}
}