Datasets:

tarekmasryo
/

global-ev-infra-dataset

pretty_name: Global EV Charging Stations & EV Models (2025)
license: cc-by-4.0
language:
  - en
size_categories:
  - 100K<n<1M
task_categories:
  - tabular-regression
  - tabular-classification
tags:
  - electric-vehicles
  - ev-charging
  - charging-stations
  - mobility
  - geospatial
  - global
  - tabular
  - dataset
  - eda
configs:
  - config_name: stations
    data_files:
      - split: train
        path: data/charging_station.csv
  - config_name: stations_ml
    data_files:
      - split: train
        path: data/charging_station_ml.csv
  - config_name: country_summary
    data_files:
      - split: train
        path: data/country_summary.csv
  - config_name: world_summary
    data_files:
      - split: train
        path: data/world_summary.csv
  - config_name: ev_models
    data_files:
      - split: train
        path: data/ev_models.csv

🌍 Global EV Charging Stations & EV Models (2025)

Author: Tarek Masryo
License: CC BY 4.0
Version: v1.0 (2025-09-15)

A clean, analysis-ready snapshot of global EV infrastructure:

Main stations table: 242,417 rows (charging sites)
Companion summaries: country + world rollups
EV models table for enrichment

📦 What’s inside (files)

All CSVs live under data/:

data/charging_station.csv — charging stations (main table)
data/charging_station_ml.csv — ML-oriented derived table (compact / engineered signals)
data/country_summary.csv — per-country rollup (counts + fast-share)
data/world_summary.csv — extended rollup (counts + power stats + fast/ultra flags)
data/ev_models.csv — EV model specs (make/model/variant + metadata)

Additional repo files:

OCM_CC_BY_4.0.txt — Open Charge Map attribution text
CHANGELOG.md, LICENSE

🧩 Why configs?

This repo includes multiple CSVs with different schemas.
Configs make the Hub viewer stable and let you load each table explicitly via load_dataset(repo_id, "<config>").

🚀 Quick start

from datasets import load_dataset, get_dataset_config_names

repo_id = "tarekmasryo/global-ev-infra-dataset"
print(get_dataset_config_names(repo_id))

# Stations
stations = load_dataset(repo_id, "stations")["train"].to_pandas()

# Summaries
country = load_dataset(repo_id, "country_summary")["train"].to_pandas()
world   = load_dataset(repo_id, "world_summary")["train"].to_pandas()

# EV models
models  = load_dataset(repo_id, "ev_models")["train"].to_pandas()

print(stations.shape, country.shape, world.shape, models.shape)

Tip: load_dataset(repo_id) will load the first config (stations) if you omit the config name.

📚 Data dictionary

`charging_station.csv` (stations table)

Typical columns include:

id, name
city, state_province, country_code
latitude, longitude
ports, power_kw
power_class, is_fast_dc

`country_summary.csv` (country rollup)

Columns:

country_code, country
station_count, port_count
fast_station_share, fast_port_share

`world_summary.csv` (extended rollup)

Columns (includes country summary + extra indicators):

country_code, country
station_count, port_count
fast_station_count, fast_port_count
fast_station_share, fast_port_share
max_power_kw, median_power_kw
dc_fast_station_count, dc_ultra_station_count
has_fast_dc, has_ultra_dc

`ev_models.csv` (EV models)

Columns:

make, model, variant
powertrain, segment, body_style
first_year, origin_country, market_regions

🎯 Suggested uses

Compare charging coverage across countries/regions
Fast-DC vs slow infrastructure analysis
Geospatial dashboards & planning
Enrich infra analytics with EV model metadata

📜 License & attribution

Dataset packaging: CC BY 4.0 — attribution required