Data Is Plural

... is a weekly newsletter of useful/curious datasets.

2023.06.14 edition

Canadian wildfires, climate finance, Pennsylvania competency cases, ad-targeting labels, and food defect thresholds.

Canadian wildfires. The Canadian Wildland Fire Information System monitors wildfires and fire-conducive conditions across the country. It publishes daily maps of its Fire Weather Index, satellite-detected hotspots, and fire intensity predictions, among other analyses. Its Canadian National Fire Database provides the perimeters, dates, sizes, and other metadata for tens of thousands of fires, spanning decades. In the agency’s datamart, you can find additional downloads of tabular and geospatial data, including burned-area statistics, active fires, and current conditions. Related: The Canadian Interagency Forest Fire Centre’s dashboard of active fires. Previously: The Global Wildfire Information System (DIP 2022.07.27). [h/t Michael Nolan]

Climate finance. “Rich nations say they’re spending billions to fight climate change” in developing countries, notes a recent Reuters investigation, which found some of this climate finance going to “questionable” projects, “including a coal plant, a hotel and chocolate shops.” To help find those examples, Reuters teamed up with Big Local News to extract “43,844 records [from the] Fifth, Fourth, and Third Biennial reports that developed countries submitted to the U.N. Climate Change secretariat.” Each of the records represents a reported contribution, listing the recipient country/region, program funded, funding amount, sector, status, and more. Read more: “How you can use climate finance data we collected for local stories around the globe” (Big Local News).

Pennsylvania competency cases. Earlier this year, Spotlight PA and the Pittsburgh Institute for Nonprofit Journalism published an investigation into Pennsylvania’s competency system, which is “supposed to protect people with mental health issues from prosecution if they cannot understand the legal system and cannot aid in their own defense.” The team has also published the data it received from the state’s court administrators, with information about 697 cases that involved competency proceedings between 2018 to mid-2022. The records, available for 23 counties, indicate each case’s court, filing date, charges, disposition, sentencing, attorneys, plus the type and timing of 1,400+ competency-related docket filings. [h/t Matt Dempsey]

Ad-targeting labels. The Markup has obtained, analyzed, and published a spreadsheet of 650,000+ ad-targetable “audience segments” (described by labels such as “Depression Propensity - Reach Tier 2”) and their data suppliers. The spreadsheet — flagged to The Markup by privacy researcher Wolfie Christl — was, until recently, linked from the website of a Microsoft-owned ad platform. “The Markup found thousands of rows in the file that indicate sensitive audience groupings,” including those related to medical issues, race/ethnicity, political activity, addiction, and more. [h/t Sharon Machlis]

Food defect thresholds. “How many rat hairs in your macaroni before the FDA considers it adulterated?” asks Cody Winchester, who has found the answer in the Food and Drug Administration’s Food Defect Levels Handbook. That guide’s “action levels” table lists 180 criteria for 111 commodities, indicating the “maximum levels of natural or unavoidable defects in foods for human use that present no health hazard.” Winchester has converted the table into a JSON file. The answer: An “average of 4.5 rodent hairs or more per 225 grams in 6 or more subsamples.”