Insights

Read the pilot workforce as a market, and show one pilot where they stand among their peers.

Insights come in two reads. A partner with insights.read reads the Workforce: what the pilots of MyPilotIndex look like as a market, in aggregates. A pilot app with a grant that holds insights.position reads where one pilot stands against that pilot’s cohort.

​

Read the Workforce

curl "https://{api-host}/v1/partner/insights/workforce/overview?country=BR" \
  -H "Authorization: Bearer $ACCESS_TOKEN"

{
  "schema_version": 1,
  "screen": "overview",
  "as_of": "2026-11-05",
  "generated_at": "2026-11-06T04:50:00Z",
  "cohort": { "status": "ok", "denominator": "cohort" },
  "privacy": { "cohort_floor": 20, "cell_floor": 5 },
  "sections": {
    "composition": {
      "status": "ok",
      "buckets": [{ "key": "major", "share_pct": 18, "suppressed": false }],
      "unpublished_pct": 2
    }
  }
}

There are six screens: overview, geography, qualifications, aircraft, score and intent. Each takes the same filters a member’s screen takes, at most three at a time. The API reference lists them. An unknown filter, a repeated one, or a value outside the vocabulary answers 400 invalid_request, and a name outside the six screens answers 404 not_found.

​

How the figures are protected

The server applies the floors, so your code never has to:

• A cohort under the floor answers status: "insufficient" and no section.
• A cell under its floor is suppressed, and the share it held shows in unpublished_pct.
• A filter that would leave too few pilots is dropped, and the answer says so in filters.broadened.

No answer carries a count of pilots. A cohort size is a band, never a number. A publication that has never run answers only schema_version, screen, generated_at and cohort, with status not_ready.

​

Show a pilot their position

curl https://{api-host}/v1/partner/grants/{grant_id}/workforce-position \
  -H "Authorization: Bearer $ACCESS_TOKEN"

{
  "schema_version": 1,
  "as_of": "2026-11-05",
  "generated_at": "2026-11-06T09:12:00Z",
  "privacy": { "cohort_floor": 20, "cell_floor": 5 },
  "cohort": { "status": "ok", "level": 2, "n_band": "250_499" },
  "metrics": {
    "total_hours": { "status": "ok", "percentile": 62, "cohort_median": 1180 },
    "score": { "status": "ok", "percentile": 71, "cohort_median": 431.5 },
    "type_rating_count": { "status": "missing" }
  }
}

The six metrics are total_hours, pic_hours, score, type_rating_count, years_experience and international_hours. Each has a status of ok, missing or insufficient, and percentile and cohort_median appear only with ok.

cohort.level says which rung of the cohort answered, from 1, the most specific, to 4, everyone. The answer names the pilot nowhere.