Is this data not free from the government?

Yes. The source files are public and we link to them. You pay for the cleaning, the joins, the documented schema and the refresh, not for the data itself.

As Parquet or CSV files, by download link or to your own cloud bucket. The same tables load into Snowflake with the scripts we include.

Can I see the data before paying?

Yes. Ask for a sample and we send every result for the 35 foods in the fruits and fruit juices group, 180,231 rows across all 30 fiscal years, plus the full data dictionary.

Why are FY2021 and FY2022 missing?

FDA publishes those years only in its TDSi dashboard, behind a terms banner that has to be accepted. This package uses only files that can be downloaded without accepting terms. Pesticide results for FY2018 to FY2020 are listed as coming soon on fda.gov, with no file linked.

Does it say whether a food is safe?

No. It holds measured levels, not a safety verdict, and it has no regulatory limits or exposure estimates. FDA changed its element methods in 2014, so part of any later rise in detections may reflect lower detection limits.

Does it include personal data?

No. Every row is a laboratory result for a food sample. No row names a person, an address or a store.

The results are US government works. FDA’s website policies say fda.gov content is in the public domain unless otherwise noted. FDA does not endorse this product.

Available now

Lead, arsenic, cadmium, pesticide residues and radionuclides measured in US foods

Every FDA Total Diet Study results file that can be downloaded, 83 files with different layouts, units and non-detect conventions, in one typed results table: 1,614,597 lab results for 477 foods and 469 analytes, FY1991 to FY2020.

Request a sample See the full catalog

Coverage: Elements, FY1991 to FY2020: 440,860 results for 31 analytes, including lead, cadmium, total and inorganic arsenic, mercury and nutrient minerals. Pesticides and industrial chemicals, FY1991 to FY2017: 1,143,844 results for 431 residues. Radionuclides, FY1992 to FY2020: 29,893 results for 6 radionuclides. 381,102 results have a detectable level. Not included: FY2021 and FY2022 for every group, and pesticide results for FY2018 to FY2020.

Why pay for public data

FDA publishes the results as 83 text files in 10 column layouts, with units, analyte spellings and non-detect conventions that change between eras. This package puts them in one table, converts every concentration to one standard unit with the published value kept beside it, stores non-detects as NULL with the limit they fall below, and merges 802 analyte spellings into 469 analytes.

How it is built

1
Pull every Total Diet Study results file FDA offers for download on fda.gov: 83 text files for elements, pesticides and industrial chemicals, and radionuclides.
2
Convert concentrations to mg/kg or Bq/kg, apply one documented non-detect rule per file family in RESULT_STATUS, and assign each food to a food group. Groups for the 120 foods not in FDA’s food and analyte matrix are assigned by QuestionPunk and marked as such.
3
Cross-check FY2018 to FY2020 element results against FDA’s own summary workbook: 6,449 of 6,449 food and element rows match. Rebuilt annually when FDA posts new downloadable results.

What is in each record

Field	Notes
RESULT_ID, FOOD_NUMBER, FOOD_DESCRIPTION, ANALYTE_ID, ANALYTE_NAME	RESULT: one row per laboratory result
FISCAL_YEAR, COLLECTION_ID, MARKET_BASKET, SEASON, REGION	REGION is published for FY2018 on only
CONCENTRATION, UNIT, RESULT_STATUS, DETECTED	Normalized to mg/kg or Bq/kg. Non-detects are NULL
LOD, LOQ, REPORTING_LIMIT, MDC, UNCERTAINTY	Detection limits and radionuclide uncertainty as published
CONCENTRATION_PUBLISHED, UNIT_PUBLISHED, FLAG_PUBLISHED, ANALYTE_NAME_PUBLISHED	The values exactly as FDA printed them
NONDETECTS_LISTED	FALSE for the 1991 to 2012 pesticide files, which list only the residues found
METHOD, DETERMINATION_METHOD, INSTRUMENT, SOURCE_FILE, SOURCE_LINE	Method fields, and the source file and line of every result
FOOD	477 foods with FOOD_GROUP and the description in each food list era
ANALYTE	469 analytes with ANALYTE_CLASS, STANDARD_UNIT, detection counts and every published spelling
COLLECTION, METHOD_CODE, DATA_DICTIONARY	130 market baskets and collections, 49 pesticide method codes and one row for each of the 106 columns

Who uses it

Food and ingredient quality teams comparing their own test results with FDA levels
Dietary exposure and food safety researchers
Journalists and advocacy groups tracking heavy metals and residues in food

Pricing

$499 a year, including each annual refresh.

No personal contact data is collected or sold in any QuestionPunk dataset.

Questions

Is this data not free from the government?: Yes. The source files are public and we link to them. You pay for the cleaning, the joins, the documented schema and the refresh, not for the data itself.
How is it delivered?: As Parquet or CSV files, by download link or to your own cloud bucket. The same tables load into Snowflake with the scripts we include.
Can I see the data before paying?: Yes. Ask for a sample and we send every result for the 35 foods in the fruits and fruit juices group, 180,231 rows across all 30 fiscal years, plus the full data dictionary.
Why are FY2021 and FY2022 missing?: FDA publishes those years only in its TDSi dashboard, behind a terms banner that has to be accepted. This package uses only files that can be downloaded without accepting terms. Pesticide results for FY2018 to FY2020 are listed as coming soon on fda.gov, with no file linked.
Does it say whether a food is safe?: No. It holds measured levels, not a safety verdict, and it has no regulatory limits or exposure estimates. FDA changed its element methods in 2014, so part of any later rise in detections may reflect lower detection limits.
Does it include personal data?: No. Every row is a laboratory result for a food sample. No row names a person, an address or a store.
What is the license?: The results are US government works. FDA’s website policies say fda.gov content is in the public domain unless otherwise noted. FDA does not endorse this product.

Tell us what you need

Geography, size and fields. A two-line email is enough, and a person replies.

Request a sample