This repository contains code and data for comparing spatial-temporal knowledge graphs (STKG) on various spatial downstream tasks. The comparisons include use cases like AirBnB and Wildfire.
- Python: Ensure you have Python installed. Recommended version is 3.11.6, which have been used for the scripts.
- R: R is required for some parts of the data analysis. Scripts are tested with R 4.2.3
If dependencies are not installed via Conda, make sure to fulfill the requirements for the different packages
Set up the Python environment:
conda env create -f environment.yml
conda activate stkg_comparisons
Set up the R environment:
conda env create -f r-environment.yml
conda activate r-stkg_comparisons
Overall we differentiate between two datasets, which we use for the overall evaluation: The AirBnB dataset as well as the Wildfire dataset. In the following, the necessary script is linked
In order to run the experiments, please have a look at the shell file containing the script to run the experiments:
The datasets needed to run the scripts can be found under the following SharePoint page
The first data preparation needs to be run with R. To have all functionalities of R, it is recommended to run R in interactive mode
In order to run the experiments, please have a look at the shell file containing the script to run the experiments:
The datasets needed to run the scripts can be found under the following OneDrive link