This repository provides a static, file-based snapshot of ACN-Data, an open electric vehicle charging dataset collected from sites at Caltech, JPL, and nearby office facilities.
The data in this snapshot covers charging activity from 2018 through 2020 and is organized for direct download, archival use, and reproducible offline analysis.
- 85,877 compressed charging-session time-series files in
.csv.gzformat. - Site folders for Caltech, JPL, and office charging locations.
- A
session data/directory reserved for session-level exports.
Current time-series coverage:
| Site group | Location | Files |
|---|---|---|
office_01 |
Parking_Lot_01 |
1,474 |
caltech |
California_Garage_01 |
31,860 |
caltech |
S_Wilson_Garage_01 |
2,703 |
caltech |
California_Garage_02 |
15,589 |
caltech |
LIGO_01 |
229 |
caltech |
N_Wilson_Garage_01 |
6,299 |
jpl |
Arroyo_Garage_01 |
27,723 |
ACN-Data-Static/
|-- README.md
|-- session data/
| `-- caltech_sessions.json
`-- time series data/
|-- caltech/
| |-- California_Garage_01/
| |-- California_Garage_02/
| |-- LIGO_01/
| |-- N_Wilson_Garage_01/
| `-- S_Wilson_Garage_01/
|-- jpl/
| `-- Arroyo_Garage_01/
`-- office_01/
`-- Parking_Lot_01/
Each file under time series data/ contains the time-series measurements for a
single charging session. File names include ACN identifiers and a session start
timestamp.
The session data/caltech_sessions.json path is included for session-level
exports. In the current checkout, the file is empty; use the time-series
directories for the complete static dataset.
The time-series files are gzip-compressed CSV files. The first column is an ISO 8601 timestamp, and the remaining columns report charging measurements:
| Column | Description |
|---|---|
Charging Current (A) |
Measured charging current in amps. |
Actual Pilot (A) |
Pilot signal current limit in amps. |
Voltage (V) |
Measured voltage in volts. |
Charging State |
Charging state label reported by the system. |
Energy Delivered (kWh) |
Cumulative energy delivered during the session. |
Power (kW) |
Instantaneous charging power in kilowatts. |
Example:
gzip -cd "time series data/office_01/Parking_Lot_01/19-102-260-1635-2019-10-31T14-21-10-330888.csv.gz" | headThe files can be read directly with standard data tools. For example, with Python and pandas:
from pathlib import Path
import pandas as pd
repo = Path("ACN-Data-Static")
path = repo / "time series data" / "office_01" / "Parking_Lot_01"
file = next(path.glob("*.csv.gz"))
df = pd.read_csv(file, compression="gzip", index_col=0, parse_dates=True)
df.index.name = "timestamp"
print(df.head())Because this repository is data-only, it does not require a package installation step. Downstream analysis code should pin its own dependencies for reproducibility.
- Official ACN-Data site: https://ev.caltech.edu/dataset
- GMM models for ACN-Data: https://github.com/tongxin-li/GMM-for-ACN-Data
- EV charging gym environment: https://github.com/tongxin-li/gym-EV
- 2024-07-05: Added the
session data/caltech_sessions.jsonexport path for Caltech session data.
If you use ACN-Data in published work, please cite the original dataset paper so others can find and use the dataset:
@inproceedings{lee_acndata_2019,
author = {Lee, Zachary J. and Li, Tongxin and Low, Steven H.},
title = {{ACN-Data}: Analysis and Applications of an Open EV Charging Dataset},
booktitle = {Proceedings of the Tenth ACM International Conference on Future Energy Systems},
series = {e-Energy '19},
year = {2019},
month = jun,
location = {Phoenix, Arizona}
}This repository does not currently include a standalone license file. Please refer to the terms on the official ACN-Data site before redistribution, and cite the dataset when using it in research or published analysis.