---
configs:
- config_name: companies
  data_files:
  - split: train
    path: data/companies/parquet/*.parquet
- config_name: locations
  data_files:
  - split: train
    path: data/locations/parquet/*.parquet
- config_name: people
  data_files:
  - split: train
    path: data/people/parquet/*.parquet
---

# OpenData Consortium

Three open datasets exported from the OpenData Consortium data platform.

| Config | Description | Primary format |
|---|---|---|
| `companies` | ~103M global companies with firmographic attributes | Parquet |
| `locations` | ~273M business locations with address and geo data | Parquet |
| `people` | ~101M business contacts linked to companies | Parquet |

## Usage

```python
from datasets import load_dataset

companies = load_dataset("OpenDataFoundation/opendata", "companies")
locations = load_dataset("OpenDataFoundation/opendata", "locations")
people    = load_dataset("OpenDataFoundation/opendata", "people")
```

## Formats

Each entity is available in three formats under `data/{entity}/{format}/`:
- **parquet** — primary, gzip-compressed
- **csv** — gzip-compressed, with header row
- **json** — newline-delimited JSON, gzip-compressed