Data Is Plural

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

2016.01.20 edition

Contaminated water, slave ships, campaign ads, Amazon food reviews, and one hyper-quantified human.

Flint water samples. Researchers from Virginia Tech have joined forces with Flint, Mich., residents to sample the city’s lead-tainted water supply. In December, the researchers posted the results of 271 samples, which indicated high levels of lead contamination. The most extreme sample found a lead concentration of 158 parts per billion — 10 times higher than the EPA’s “action level.” Related: The New York Times + The Washington Post have used the data.

The transatlantic slave trade. Slate Magazine’s “The Atlantic Slave Trade in Two Minutes” — recently named a multimedia finalist for the American Society of Magazine Editors’ annual awards — tracks 20,528 transatlantic voyages over 315 years. The information comes via, which provides searchable, downloadable records of ships’ and captains’ names, regions where slaves were purchased and sent, and more.

Campaign ad purchases. The FCC requires broadcasters to keep records of “all requests for broadcast time made by or on behalf of a candidate for public office.” With the help of volunteers, Political Ad Sleuth gathers those records and enters them into a searchable, downloadable database. Note: Due, in part, to the difficulty of transcribing the (non-standardized) records, the information in the database is incomplete.

568,454 reviews of “fine foods” on Amazon. In 2013, Stanford University researchers published a paper examining how people’s tastes “change and evolve over time.” They drew, in part, on a dataset containing 13 years of Amazon reviews of gourmet foods. (Note: Not all foods were intended for humans.) The dataset comes in a slightly unconventional format; here’s a Python script to convert it to a TSV file. [h/t Kaggle]

One hyper-quantified human. Last month, Nature Communications published a study of the “long-term neural and physiological phenotyping of a single human.” That human? Study co-author Russell A. Poldrack, “a right-handed Caucasian male, aged 45 years at the onset of the study.” The 18 months of results — tracking brain connections, food consumption, stress levels, and much more — are available to download and explore. [h/t Sune Lehmann]