lmoncla/lyon-velov-bike-sharing-dataset

---
license: etalab-2.0
region: eu
tags:
- mobility
- lyon
- bike-sharing
- smart-city
dataset_info:
features:
- name: station_id
dtype: int64
- name: status
dtype: string
- name: capacity
dtype: int16
- name: bikes_total
dtype: int8
- name: bikes_mechanical
dtype: int8
- name: bikes_electric
dtype: int8
- name: stands_free
dtype: int16
- name: departures
dtype: int8
- name: arrivals
dtype: int8
- name: horodate
dtype: timestamp[ns, tz=UTC]
splits:
- name: train
num_bytes: 728514475
num_examples: 21717657
download_size: 116719084
dataset_size: 728514475
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

# Lyon velo'v Bike Sharing Dataset

## Dataset Description

This dataset provides an aggregated view of the Vélo'v bike-sharing station activities in the Lyon Metropolitan area (France) for 2023, 2024, and 2025. Unlike the raw real-time data, this dataset has been preprocessed to offer a 30-minute temporal granularity, including calculated flows (bike arrivals and departures).

**Note on Temporal Coverage:** The current release ensures a continuous and comprehensive record for all of **2023**, **2024**, and **2025**.

* Original Source: [Métropole de Lyon - Open Data Portal](https://data.grandlyon.com/portail/fr/jeux-de-donnees/stations-velo-v-de-la-metropole-de-lyon---disponibilites-temps-reel/info).
* License: [Etalab Open License](https://www.etalab.gouv.fr/wp-content/uploads/2017/04/ETALAB-Licence-Ouverte-v2.0.pdf).

## Usage

You can easily load this dataset using the Hugging Face `datasets` library:

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("lmoncla/lyon-velov-bike-sharing-dataset", split="train")

# Convert to a Pandas DataFrame for analysis
df = dataset.to_pandas()

# Display the first few rows
print(df.head())
```

## Preprocessing Details

The data cleaning and transformation were performed using Python scripts. Key steps include:

* **Normalization**: Conversion of raw JSON snapshots into a structured tabular format.
* **Flow Calculation**: Using the difference in total bike counts to identify departures and arrivals for each station.
* **Resampling**: Time-series aggregation into 30-minute buckets, preserving the last known capacity and summing up movements.

## Dataset Schema

* `station_id`: Unique identifier for the bike station.
* `horodate`: Timestamp marking the end of the 30-minute interval.
* `capacity`: Total number of docks available at the station.
* `bikes_total`: Total number of bikes available at the end of the interval.
* `bikes_mechanical` / `bikes_electric`: Breakdown by bike type.
* `departures`: Cumulative sum of bike departures within the 30-minute window.
* `arrivals`: Cumulative sum of bike arrivals within the 30-minute window.

## Citation and Attribution

Original data produced by Métropole de Lyon.

> Note: This dataset is a derivative work created for analysis and academic purposes.

If you use this dataset in your research or project, please cite it as follows:

```bibtex
@misc{lyon-velov-bike-sharing-dataset-2026,
author = {Moncla, Ludovic},
title = {Lyon velo'v Bike Sharing Dataset},
year = {2026},
publisher = {Hugging Face},
journal = {Hugging Face Datasets},
howpublished = {\url{[https://huggingface.co/datasets/lmoncla/lyon-velov-bike-sharing-dataset](https://huggingface.co/datasets/lmoncla/lyon-velov-bike-sharing-dataset)}}
}