Scraped Dataset QA Preflight

Scraped Dataset QA Preflight checks sampled scraper output before delivery or downstream import. It focuses on source-specific, machine-detectable signals: missing fields, duplicate key rows, format-looking problems, mixed value types, sensitive-looking pattern counts, and lightweight schema contract mismatches.

It can read a dataset ID, resolve an Actor run ID to its default dataset, or analyze submitted JSON/CSV data. API tokens and raw row values are not stored in the report or Evidence Pack.

Presets

• googleMaps: name, address, phone, website, place URL, category, and hours signals
• leadList: email, company, person, phone, domain, and source URL signals
• socialProfile: username, profile URL, follower/post counts, bio, and external URL signals
• generic: field profile, duplicates, formats, sensitive-pattern preview, and optional contract checks

Boundaries

This is a sampled dataset-output preflight. It does not independently check source pages, confirm complete crawl coverage, repair scraper output, judge legal suitability, validate live email inboxes, or predict downstream outcomes.

Billing

Recommended launch pricing:

• scraped-dataset-qa-report-generated: USD 0.50 per generated QA report
• scraped-dataset-row-sampled: USD 0.01 per sampled row

Invalid input and dataset load/inspect failures are not charged.

Related check

If the same scraper runs repeatedly, use Dataset Drift Monitor to compare the current aggregate dataset shape with its latest compact baseline.

Local Example

$PYTHONPATH=src python -m scraped_dataset_qa_preflight.cli examples/sample-input.json --json report.json --md report.md