DC's War on Duplicate Street Imagery: How the Capital Stacks Up Against London, Seoul, and São Paulo

Washington DC's Office of the Chief Technology Officer has been quietly working through a backlog of duplicate photographic records embedded in the city's municipal GIS mapping platform — a problem that sounds bureaucratic until you realize it's slowing down everything from pothole repair requests to emergency dispatch routing. The redundant imagery, accumulated across multiple scanning cycles dating back to at least 2019, clogs the District's 311 service infrastructure and complicates updates to the city's open data portal at opendata.dc.gov.

The timing matters. With the Trump administration's DOGE-driven restructuring cutting into federal contracts that many DC agencies depend on for technology support, the District is absorbing costs it had previously shared with federal partners. Mayor Muriel Bowser's technology office is now handling deduplication largely through internal staff rather than the contracted vendors that managed earlier data-cleaning cycles. That shift has slowed the process, according to public procurement records posted to the DC Council's oversight database this spring.

What Deduplication Looks Like on the Ground

The practical consequences show up in specific neighborhoods. In NoMa, where rapid development since 2020 has meant near-constant street-level changes, the mapping platform still serves up outdated imagery alongside current scans for stretches of Florida Avenue NE and New York Avenue NE. Community organizations in Anacostia, including the Anacostia Community Development Corporation, have flagged similar issues when using city data tools to document infrastructure needs — duplicate records create conflicting timestamps that make it harder to demonstrate when a problem first appeared.

The DC Office of Planning confirmed in a March 2026 budget presentation to the Council that it allocated $1.2 million in fiscal year 2026 for data infrastructure maintenance, a line item that covers deduplication alongside broader database hygiene. That figure is roughly flat compared with FY2025, even as the volume of imagery in the system has grown by an estimated 30 percent since the city expanded its street-scanning program in 2023.

How Other Cities Are Handling the Same Problem

London's Ordnance Survey and Transport for London jointly implemented an automated deduplication pipeline in 2024, using hash-comparison algorithms to flag identical or near-identical images before they enter the live database. The system reduced redundant records in TfL's street asset database by roughly 40 percent within six months of deployment, according to a published TfL technical report from January 2025. Seoul's Smart City Division in South Korea took a different route, contracting with a domestic AI firm in 2023 to scrub its urban imagery archive ahead of a citywide infrastructure audit — the project covered approximately 11,000 kilometers of mapped roadway.

São Paulo's municipal data agency, IPT, has faced challenges more comparable to DC's: an aging platform, inconsistent scanning schedules across different city departments, and budget constraints that pushed a planned overhaul from 2024 into 2026. São Paulo's experience is instructive because it shows what happens when deduplication is deferred — downstream services including the city's digital permitting system for construction projects ground to a halt twice last year when conflicting image records created validation errors in automated workflows.

DC is closer to São Paulo's situation than London's right now. The District lacks an automated pre-ingestion filter. Images are scanned, uploaded, and then reviewed manually in batches — a workflow that made sense when the platform was smaller but has not scaled to current data volumes. The OCTO has floated a pilot program to introduce algorithmic flagging, but as of June 2026 no contract had been awarded, according to procurement listings on the DC government's contracting portal.

For residents and local developers who rely on the open data portal — whether to build civic apps, file accurate 311 reports, or cross-reference permit histories on H Street NE or Martin Luther King Jr. Avenue SE — the practical advice is straightforward: treat any imagery timestamped before January 2024 as potentially superseded by a newer scan that may not yet be correctly indexed. The OCTO's data stewardship team does accept error reports through the opendata.dc.gov feedback form, and submissions from verified civic organizations are prioritized in the review queue. The automated fix is coming, eventually. Until then, the workaround is manual.