AIPARK-Benchmark-Berlin

This repository contains the code and data required to reproduce the results from the blog post Automating quality tests for map data to solve urban parking . You can find the full article on the Benchmark of AIPARK Predictive Parking Data in Berlin here.

Requirements

In order to run the project, you need to have Python 3.6 installed. You will need to have these libraries installed:

• numpy
• requests
• urllib3

You can install new Python libraries using the pip command (https://pypi.org/project/pip/). To install them all at once, run:

pip3 install numpy requests urllib3

Setup

1. To run this project, clone or download the repository and cd into the directory "AIPARK-Benchmark-Berlin".

2. Next, go to https://studio.aipark.io/sign-up to obtain your personal API key. It's free and done in a minute.

3. Open compare_predictions.py with a text editor of your choice and put in your API key at the bottom of the file. Then you are all set.

if __name__ == "__main__":
    api = AIPARK_API(apikey="insert-your-api-key-here")
    run_comparison()

4. Now, run compare_predictions.py on your terminal.

python3 compare_predictions.py

5. The script now imports the ground truth data from ground_truth.csv and fetches predictive parking data from the AIPARK API. Finally, the prediction metric as defined in the blog post gets applied and the results are printed out on the command line.