Geography on Big Muddy

Vandalism of OpenStreetMap

Mon, 23 Mar 2026 17:26:00 -0400

OpenStreetMap (OSM) is an open, community-driven map database powering countless apps and services and used by organizations including Amazon, Apple, Microsoft, Uber, Mapbox, and Wikimedia. In short, it is foundational infrastructure for the web. For regions with active communities (particularly in Europe), OSM is often noted for the superiority of its data on features such as cycling routes, hiking trails, and footpaths.

The Wikipedia article for OpenStreetMap documents several instances of data vandalism, which OSM is vulnerable to as a crowdsourced project. Three incidents stood out:

In 2012, Google fired two “rogue contractors” for vandalizing the OSM database, intentionally adding false data such as reversing the direction of one-way streets.
In 2018, a vandal made several viciously antisemitic edits to place names around New York City. While quickly reverted at the source, these changes nonetheless propagated into downstream applications pulling data from MapBox, such as Zillow, Snapchat, Citibike, and Wikipedia.
Users of the mobile game Pokémon GO regularly vandalize the OSM database underlying the game to gain a gameplay advantage, although the authors of the research article on this subject note this vandalism tends to be transitory rather than sustained.

Side note: I was amused to note how strong Google’s regional results bias is for “OSM”—the entire first page is taken up by results related to the Orchestre symphonique de Montréal.

For map nerds only: An atlas of world history

Sat, 21 Mar 2026 22:39:00 -0400

I am sharing today TimeMap.org: an atlas of regions, rulers, people, and battles throughout history. Thoroughly enjoyable to swipe through, especially for connoisseurs of the map game genre.

^{Hat tip to agilek on Hacker News.}

geoBoundaries: An open database of political administrative boundaries

Fri, 13 Mar 2026 17:05:00 -0400

Today I discovered geoBoundaries, a CC BY 4.0-licensed database of political administrative boundaries covering the entire world. It is notable for its high level of detail, going from ADM0 (country), ADM1 (states/provinces), ADM2 (counties/departments or municipalities), to ADM3 (municipalities or sub-municipalities) for many countries. My go-to source for world map files is Natural Earth, which is limited to ADM0 and ADM1 but is in the public domain. Natural Earth also includes some physical geography like water and bathymetry, while geoBoundaries is focused solely on political administrative boundaries. Both datasets deal with disputed boundaries, which is an endless source of tension in the Natural Earth GitHub.

An R package for retrieving data from geoBoundaries, geobounds, was released in February. A similar package for Natural Earth, rnaturalearth, has long been maintained by rOpenSci.