UrbanEV v1.0.0 - First Stable Dataset & Forecasting Toolkit Release 🚀

🚀 UrbanEV v1.0.0 — Stable Release

We are excited to announce the first stable release of UrbanEV 🎉

UrbanEV is a large-scale open dataset for electric vehicle (EV) charging behavior analysis and prediction, collected from public EV charging stations in Shenzhen, China. This release makes publicly available the full dataset, preprocessing scripts, and model training code for both traditional and deep learning-based spatiotemporal forecasting.

---

📂 Primary Data Repositories

You can find the complete dataset (raw + cleaned) at the following locations:

• 📂 Dryad: https://doi.org/10.5061/dryad.np5hqc04z
• 📁 Google Drive (latest version): Google Drive Folder

---

🌍 Dataset Highlights

📍 General Info

• Location: Shenzhen, China
• Collection Period: Sep 1, 2022 – Feb 28, 2023

📦 Raw Dataset

• 🚗 Stations: 1,682 public EV charging stations
• ⚡ Piles: 24,798 EV charging piles
• 🗺️ Zones: 331 traffic zones
• 📊 Granularity: 5-minute resolution
• 📁 Storage: Dryad, Google Drive

✅ Cleaned Dataset (used in benchmark)

• 🚗 Stations: 1,362 public EV charging stations
• ⚡ Piles: 17,532 EV charging piles
• 🗺️ Zones: 275 traffic zones
• 📊 Granularity: 5-minute and 1-hour resolution
• 📁 Storage:
  • Hourly benchmark data: this repository’s data/ folder
  • Full cleaned data: Dryad, Google Drive

---

🛠️ What's Included

• ✅ Cleaned zone-level dataset with rich spatiotemporal features
• ✅ Raw station-level data for custom processing
• ✅ Code for:
  • Baseline models (AR, ARIMA, FCNN, LSTM, GCN, ASTGCN)
  • Transformer-based forecasting models
• ✅ Ready-to-run scripts for training and evaluation
• ✅ Support for environment setup on Windows/Linux

---

📚 Citations

If UrbanEV helps your work, please consider citing the following papers:

• UrbanEV Benchmark Dataset: Scientific Data, 2025
• Physics-informed GNN: IEEE T-ITS, 2023
• Electricity Price Sensitivity Study: Sustainable Cities and Society, 2024
• ChatEV (NLP-based Forecasting): TRD, 2024

---

📦 Installation & Usage

See README for full setup instructions, model training commands, and environment configuration.

---

📩 Contact

Maintainers:

• Han Li – [email protected]
• Haohao Qu – [email protected]

---

This project is released under the CC0 1.0 Universal License.