Spatial-Economic Analysis for Optimal EV Charging Station Placement.

Group 7, Data Science 2023
Dashboard · Presentation · Video . Report

Table of Contents

1. About The Project
2. Folder structure
3. Data and Usage
4. Dashboard
5. Team Members
6. Acknowledgments
7. Paper and Citation
8. Contact

About The Project

Spatial-Economic Analysis for Optimal EV Charging Station Placement using Machine Learning.

The global transportation sector has been facing serious issues due to the use of internal combustion engines (ICEs), including their contribution to global warming. The exhaustion of fossil fuels and greenhouse gas emissions, particularly carbon dioxide, have raised concerns about climate change and global warming. To combat these problems, there is growing interest in transitioning from ICEs to cleaner and more sustainable alternatives, such as electric vehicles (EVs). EVs, running on electricity produced from renewable sources, offer the advantage of zero tailpipe emissions and can significantly reduce greenhouse gas emissions in the transportation industry. The transition to EVs has gained momentum, driven by declining battery prices and advancing charging methods. However, the widespread adoption of EVs requires an expansion of public charging infrastructure, necessitating the optimal placement of EV charging stations (EVCS).

In our work we focused on following task:

• Geographical and socio-economic variables can serve as indicators or proxies for understanding the demand for EV charging stations.
• The project aims to create an exhaustive dataset considering socio-demographic features of Germany and solve the problem of optimal EVCS placement using classical ML algorithms.
• The study evaluates different ML models and compares their performance to identify the most suitable approach.
• The findings contribute to the understanding of optimal EVCS placement and facilitate automated decision-making in expanding EV charging infrastructure.

Folder structure

• All the code can be found under notebook/
• The final dataset can be found under data/processed/all_city_data_with_pop.csv

├── Dashboard
│   ├── app.py
│   ├── Data
│   ├── MyMap.html
│   ├── README.md
│   └── requirements.txt
├── data
│   └── processed
│       ├── all_city_data_with_pop.csv
│       ├── berlin_data_detailed.csv
│       ├── Frankfurt_data_detailed.csv
│       ├── Kaiserslautern_data_detailed.csv
│       ├── karlsruhe_data_detailed.csv
│       ├── Mainz_data_detailed.csv
│       └── Saarbrücken_data_detailed.csv
│   └── raw
├── figures
│   ├── XX.png
├── notebooks
│   ├── CITY_NAME.ipynb
│   ├── cache
│   ├── Data_Science_Mini_project_EDA.ipynb
│   ├── EDA.ipynb
│   ├── make_data_set.ipynb
│   ├── map_images
│   │   ├── xx.jpg
│   ├── modeling.ipynb
│   ├── plots.ipynb
└── README.md

Data and Usage

Data Collection Pipeline

How replicate results

1. set path_root_dir in notebook/modeling.ipynb
2. results will be created under result directory

How to create dataset from raw data

1. Download data from here
2. Set path_to_downlaoded_data in notebooks/make_data_set.ipynb
3. run the notebook and data will be created

Dashboard

The dashboard is created with Streamlit in Python, which is an open-source framework designed to create interactive web apps. Our app is hosted on Hugging Face. The dashboard shows a map of Saarbrücken, Germany, with points-of-interest, existing EV charging stations, residential, and commercial areas marked on the map, along with the optimal location of new EV charging stations predicted by our model. The app gives the option to view different features marked on the map that were collected and used to train our model. Final results and EDA plots, for example, a pie chart of the number of different types of infrastructure in Saarbrücken, are also displayed on the dashboard to help visualize the data and evaluate overall model performance.

Data Collection Pipeline

Link to the app hosted on hugging-face

Dashboard code can be found under: Dashboard

Roadmap

• Collect and gather data for the task
• Performed EDA EDA_Notebook_1 EDA Notebook_2
• Developed Machine Learning Model
• Deployment of final prototype to Dashboard for the end user.

Team Members

1. Bahram Khan Baloch (7047281)
2. Saira Sohail Anwari (7047706)
3. Umer Butt (7024124)
4. Cicy Kuriakose Agnes (7047703)
5. Akansh Maurya (7047939)

Acknowledgments

We would like to thank Prof. Dr.-Ing. Wolfgang Maaß and tutors of Data Science course 2023 at Saarland University for giving us the opportunity to work in such a industry relevant project. We would also like to acknowledge the following repositories/organization for making this project successful.

• OpenStreetmap: (https://www.openstreetmap.org)
• Germna Federal Network Agency: (https://www.bundesnetzagentur.de)
• Humanitarian Data Exchange: (https://data.humdata.org)
• Data for Good at Meta:
• Where Should We Be Placing EV Charging Stations? (Using QGIS and PulP)

Paper and Citation

Please use the following citation for our data or methodology:

@INPROCEEDINGS{10607132,
  author={Maurya, Akansh and Agnes, Cicy Kuriakose and Baloch, Bahram Khan and Anwari, Saira Sohail and Butt, Umer},
  booktitle={2024 International Conference on Social and Sustainable Innovations in Technology and Engineering (SASI-ITE)}, 
  title={Spatial-Economic Analysis for Optimal Electric Vehicle Charging Station Placement}, 
  year={2024},
  volume={},
  number={},
  pages={339-344},
  keywords={Technological innovation;Local government;Pipelines;Charging stations;Data collection;Electric vehicle charging;Data models;Machine Learning;Open street map;Germany;EV charging station placement},
  doi={10.1109/SASI-ITE58663.2024.00071}}

Contact

If dashboard is not working, please create an new issue or email at: akanshmaurya@gmail.com