CS3 Case Study

Electric Vehicle Charging Demand Forecasting

Course DS 4002 Section 001
Date 12/2025
Contributor Beverley Appiatse
Instructor Professor Alonzi

Contents of the Repository

This repository contains all materials for the CS3 forecasting project. The structure supports clarity transparency and reproducibility.

• README.md
• Scripts
• Supplemental_Materials
• Data
• Rubric.pdf
• Data_Appendix (2).pdf

Section 1: Software and Platform

Software used

• Python 3.10
• Jupyter Notebook

Required packages

• pandas
• numpy
• matplotlib
• scikit-learn
• tensorflow or pytorch

Install using

pip install -r requirements.txt

Section 2: Repository Structure

Scripts folder

Contains all forecasting notebooks

• CHARGE_Arima.ipynb
• Charge_LSTM.ipynb
• CHARGE_TemporalCNN.ipynb

Data folder

• Data Appendix (2).pdf
• Dataset Establishment Details.md
• LA_Volume (1).csv
• LA_Volume_Cleaned (2).csv

Supplemental Materials folder

• Mastering Time Series Forecasting_ From...A to LSTM - MachineLearningMastery.com.pdf: Beginner friendly forecasting explainer
• Forcast_of_EV_Charging_Demand.pdf: EV charging domain article

Rubric.pdf

The project rubric and evaluation criteria

Hook Document.pdf

The Hook Document.pdf provides an introduction to the forecasting problem and explains the scenario context for the project. It serves as a starting point before beginning work in the Scripts folder.

Section 3: Instructions for Reproducing the Workflow

• Clone the repository

git clone https://github.com/beverleyappiatse/CS3.git

cd CS3

• Install dependencies

pip install -r requirements.txt

• Open the Scripts folder to begin the analysis

• Conduct the baseline model Open CHARGE_Arima.ipynb and follow the workflow to load data prepare timestamps construct an ARIMA model and gather predictions

• Conduct an advanced model Open either Charge_LSTM.ipynb or CHARGE_TemporalCNN.ipynb and complete the steps for building an advanced forecasting model

• Generate evaluation results Each notebook includes cells that compute RMSE MAE and plot actual versus predicted values

• Compare models Use metrics and visuals to determine which approach performs best

Section 4: Purpose of the Project

This project explores electric vehicle charging demand through a complete forecasting pipeline. The workflow includes data acquisition preparation modeling evaluation and interpretation using real world EV charging data.

Section 5: References

• Guo X Zhang Y and Li T CHARGED Global EV Charging Dataset Harvard Dataverse

• Machine Learning Mastery Time Series Forecasting Tutorial

• SpringerOpen Probabilistic Forecast of EV Charging Demand

License

This repository is authored by Beverley Appiatse under MIT License unless otherwise noted.