Data Is Plural

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

2020.08.12 edition

Coronavirus humanitarian funding, US climate change, international banking statistics, electric vehicle charging, and 44 million Reddit votes.

Coronavirus humanitarian funding. The Centre for Disaster Protection and Development Initiatives have been jointly compiling data on the billions of dollars of humanitarian and development aid that the IMF, World Bank, and other agencies have allocated in response to the COVID-19 pandemic. For each “flow” of funds, the dataset specifies the funding source, amount, approval date, purpose, and more. Related: The UN’s provides a downloadable and explorable dataset of its coronavirus–related humanitarian funding, and has built an interactive map based on these and related datasets.

US climate change. On Friday, the Washington Post released several datasets and computer scripts from its Pulitzer Prize–winning series, “2°C: Beyond the Limit,” which examined how “extreme climate change has arrived in America.” The data files — derived from NOAA’s nClimDiv and nClimGrid climate datasets — contain the annual average temperatures and seasonal temperature changes for each state and county in the contiguous US.

International banking statistics. The Bank for International Settlements, established in 1930 and run by a group of central banks, publishes a range of statistical datasets “designed to inform analysis of financial stability, international monetary spillovers and global liquidity.” The datasets are available to explore online and to download; they cover exchange rates, cross-border liabilities, consumer prices, debt service ratios, and more. [h/t Erik Gahner Larsen]

Electric vehicle charging. A team led by public policy professor Omar Asensio used a field experiment to collect data on 3,395 electric vehicle charging sessions. The dataset “contains sessions from 85 EV drivers with repeat usage at 105 stations across 25 sites at a workplace charging program”; it indicates the date and length of each session, total energy used, cost, and more.

44 million Reddit votes. On Reddit, users can choose to publicize their history of upvoting and downvoting other users’ posts. Software engineer Joey Leake recently collected and published data on more than 44 million of these votes. For each, Leake’s dataset lists the post ID, relevant subreddit, the vote’s timestamp (in most cases), and the voter’s username.