edwarddgao/open-apply-jobs · Datasets

edwarddgao/open-apply-jobs · Datasets at Hugging Face Open-Apply Jobs A daily-refreshed open dataset of active job postings sourced directly from public ATS APIs (Greenhouse, Lever, Ashby). Every record can be traced back to the hiring company's own career board. - Refresh: automated daily at 06:00 UTC - Partitioning: Hive-partitioned Parquet (`date=YYYY-MM-DD/source={ats}`) - Source code: https://github.com/edwarddgao/openapply Usage ``` from datasets import load_dataset ds = load_dataset('edwarddgao/open-apply-jobs') ``` ``` Latest snapshot only import duckdb duckdb.sql(""" SELECT title, source_slug, apply_url, locations FROM read_parquet('hf://datasets/edwarddgao/open-apply-jobs/data//.parquet', hive_partitioning=1) WHERE date = (SELECT MAX(date) FROM read_parquet( 'hf://datasets/edwarddgao/open-apply-jobs/data//.parquet', hive_partitioning=1)) AND 'Software Engineer' = ANY(string_split(title, ' ')) """).show() ``` Partition layout ``` data/ ├── date=YYYY-MM-DD/ │ ├── source=greenhouse/part-.parquet │ ├── source=lever/part-.parquet │ └── source=ashby/part-.parquet ├── date=YYYY-MM-DD/ │ └── ... ``` Each`date=` folder is a full snapshot of that day's active postings — not an incremental delta. Diff two consecutive dates to get added / removed / modified jobs. Schema | Field | Type | Notes | | --- | --- | --- | | `id` | string | `{source}:{slug}:{native_id}`— unique across the dataset | | `source` | string | `greenhouse`|`lever`|`ashby` | | `source_slug` | string | Tenant slug on that ATS (e.g.`databricks`,`spacex`) | | `title` | string | Job title | | `apply_url` | string | Canonical URL on the ATS career site | | `description_html` | string? | Full HTML description (~100% populated) | | `employment_type` | string? | e.g.`FullTime`,`Contract`,`Internship` | | `department` | string? | Free-form; ATS-dependent | | `locations` | liststring | Always a list; may be empty for fully-remote |