Client: Strelka KB (Moscow)
Location: 32 single-industry towns across Russia
Year: 2016–2017
Sector: Social Media Data · Computer Vision · Space Syntax · Research at Scale
Reading the everyday public life of 32 Russian monotowns through a million citizen-shared photos — because no formal urban data existed for these towns at all.
Russia’s monotowns — cities built around a single industry — house roughly a sixth of the national population, but almost none of them have usable planning data: no digitised street networks, no amenity surveys, no record of how public space is actually used. Strelka KB needed to understand these towns’ social life well enough to design interventions that could scale across 300+ of them.
Commissioned by Strelka KB as part of a national design-response programme to monotown decline — a live policy issue in Russia since unrest in single-industry towns (notably Pikalyovo, 2008) put the topic on the national agenda. SPIN Unit led the research, with Prof. Panu Lehtovuori, data scientist Raul Kalvo, and academic collaborators Lev Manovich (Cultural Analytics) and Daniele Quercia, Luca Maria Aiello, and Rossano Schifanella of Nokia Bell Labs (computer vision).
With no existing GIS base data, the team hand-digitised street networks from OpenStreetMap for all 32 towns and ran pedestrian betweenness analysis calibrated from an earlier SPIN Unit study for Strelka KB in Moscow. Roughly a million geolocated VKontakte photos were collected; a sample of 45,000 was manually tagged by eight trained researchers using a custom activity taxonomy (1,126 terms, built from academic literature and hashtag-frequency analysis) and a bespoke tool built for the project, cross-validated against Nokia Bell Labs’ computer-vision scene classification. Three towns received a deeper qualitative pass using Lev Manovich’s Cultural Analytics image-montage technique.
Delivered as a 33-page methodology report plus a 208-page, two-volume atlas of per-city maps and data spreads covering all 32 towns.
- Family and domestic life dominate monotown social media imagery — birthdays, weddings, school events — while café culture, restaurants, and nightlife are largely absent; commercial “third places” barely register.
- Youth cluster at two poles: the busiest central squares, or informal/marginal spaces (bridges, edge-of-estate woodland, railway land) — but rarely venture into nature even when it’s nearby.
- The founding factory is often invisible in residents’ own imagery — closed, avoided, or simply not photographed — a striking absence given it’s the reason each town exists.
- No single “monotown template” works: despite shared Soviet industrial origins, the 32 towns show highly varied morphology, history, and topography, ruling out generic, one-size-fits-all interventions.
- Using citizen-generated photos rather than researcher observation was a deliberate methodological choice to reduce professional bias — while the team was explicit that social data alone isn’t sufficient and needs triangulation.
Gave Strelka KB an evidence base — at a genuinely unprecedented 32-town scale — for scalable, cost-efficient public-space interventions across Russia’s monotowns, grounded in what residents actually do rather than planning assumptions.