How open-source investigators pinpoint locations in conflict videos using landmarks, satellite imagery, and digital forensics.
When a ten-second video emerges on social media showing an explosion, a damaged building, or military activity, its value as evidence depends entirely on one question: where was it taken? Without a precise location, the footage is anecdotal—impressive perhaps, but impossible to verify or map accurately. Geolocation is the discipline of answering that question using open-source intelligence (OSINT) techniques.
In the Ukraine conflict and Middle Eastern wars, thousands of videos are uploaded daily by soldiers, civilians, and journalists. Many contain no metadata indicating location. Yet skilled analysts can often identify the exact building, street, or military installation in a few minutes, using only what is visible on screen and freely available online tools. This process combines detective work, local knowledge, and digital forensics.
The first step is always observation. An analyst watches a video frame by frame, cataloguing everything visible: building styles, street signs, power lines, shop fronts, vehicles, weather, shadows, and terrain features. A distinctive building—a church steeple, water tower, or unusual roof structure—can anchor an investigation. A shop sign written in Cyrillic narrows the search to former Soviet space. A Ukrainian flag flying from a building confirms the country.
Landmark databases like Google Street View, Yandex Maps (which has better coverage of Ukraine and Russia), and regional mapping services become essential reference tools. An analyst might search for 'blue-roofed church near Kharkiv railway station' and scroll through Street View images until finding a match. Road signs, vehicle license plates, and architectural details visible in the video are cross-referenced against imagery collected in peacetime. A distinctive graffiti tag, a broken street lamp in a specific location, or the layout of a town square—any of these can serve as a fingerprint.
In built-up areas, distinctive landmarks are abundant and verification can be rapid. In rural areas with fewer visual references, the process becomes harder. An analyst might need to triangulate using multiple features—the direction of shadows, the type of crop in a field, power line configuration, or the shape of a distant hill.
Once an analyst has a tentative location from visual landmarks, satellite imagery confirms it. Services like Google Earth, Sentinel Hub (operated by the European Commission), and Maxar provide high-resolution satellite photographs taken at different dates. These are invaluable for two reasons: they show the ground from directly overhead, eliminating the distortion of perspective; and they reveal changes over time that correlate with known events.
A video showing a destroyed apartment building can be compared against satellite images from before and after the destruction date. If the building appears intact in imagery from March and damaged in imagery from April, and the video shows that damaged state, the timeline aligns. Multiple satellite images from different dates also help rule out false positives—if a video shows a red-roofed house beside a blue fence, and satellite imagery shows no such combination anywhere in the suspected area, the hypothesis is rejected and the search continues.
Satellite imagery has limitations. Cloud cover can obscure scenes. High-resolution imagery is not available for all areas on the same schedule. Buildings look different from directly overhead than from ground level, which can complicate matching. Nevertheless, satellite photography remains the geographic gold standard because it is objective, time-stamped, and verifiable by others.
Video files contain metadata—embedded information about when and where they were created. This metadata can be corrupted, spoofed, or stripped away, especially if footage passes through multiple social media platforms. Platforms like Twitter, TikTok, and Telegram compress and re-encode video, destroying original file information. Nevertheless, if a video is obtained directly from a source, metadata examination is a crucial first step.
Metadata analysis uses open-source tools to examine embedded GPS coordinates, timestamp information, camera settings, and file creation details. These are rarely conclusive on their own—GPS coordinates can be falsified, timestamps can be changed—but they provide initial clues. An analyst might note that a video was created on a specific date, then search for news events in candidate locations on that date to narrow possibilities.
Beyond metadata, analysts examine visual clues invisible to casual observers. Shadow angle and length can indicate the time of day and, combined with the season, help narrow location. The position of the sun creates shadows at predictable angles and times depending on latitude and time of year. A video showing long shadows consistent with early morning or late afternoon in summer at high latitude suggests a location far north. These geometric details are subtle but powerful filters when combined with other evidence.
Some geolocation challenges are solved through crowdsourcing. When an analyst publishes preliminary findings on platforms like Twitter or Reddit asking 'Does anyone recognize this street?', locals or people familiar with a region often respond. A Ukrainian living in Dnipro might instantly recognize a downtown café or intersection. A journalist based in Syria might identify a neighbourhood by distinctive graffiti or architecture. This collective intelligence accelerates identification.
Local knowledge is particularly valuable in smaller towns or neighbourhoods where Street View coverage is sparse. An analyst working on a video filmed in rural Ukraine might crowd-source through Ukrainian language forums or diaspora communities online. However, crowdsourcing carries risks: false information, deliberate misinformation, or simply well-intentioned but incorrect guesses. Responsible analysts always verify crowd-sourced claims against satellite imagery and other independent evidence before publishing results.
Community contributions have also produced detailed databases of destroyed buildings, military positions, and damage assessments organized by location and date. These crowd-sourced archives become reference libraries for future geolocations, since analysts can search whether a particular building or landmark appears in existing, already-verified footage.
A geolocation is only credible when multiple independent lines of evidence point to the same location. An analyst might identify a building by its architecture (matching Street View), confirm the location using satellite imagery (showing the same surrounding buildings and streets), note that the video timestamp aligns with known military activity in that area, and have the location verified by a local who recognizes the scene. Each piece of evidence strengthens confidence.
Responsible analysts publish not just their conclusion—'This video was filmed at coordinates X,Y'—but also their reasoning and sources. They photograph or screenshot the Street View image, the satellite image, and other references they used. They state what they are confident about and what remains uncertain. This transparency allows others to review their work, spot errors, and build on their findings. It also makes it harder to spread deliberate misinformation, since the public can evaluate the evidence.
Quality control breaks down when speed is prioritized over accuracy. During fast-moving conflicts, there is pressure to geolocate and publish findings rapidly to shape the news cycle. Responsible analysts resist this pressure, taking time to verify thoroughly even if it means publishing later than competitors. A single false geolocation—attributing an image to the wrong location or date—can spread widely and damage credibility and produce harmful misinformation.
For readers consuming conflict maps and geolocated footage, skepticism is warranted. Ask: Is the source showing their working? Are multiple pieces of evidence cited, or just a single landmark? Has the geolocation been independently verified by other analysts or news organizations? Do the building damage, weather, and lighting appear consistent with known events on the stated date?
The discipline of geolocation has become central to conflict documentation and accountability, but it is not foolproof. Mistakes happen. Deliberate deception—filmed recreations or decades-old footage misrepresented as current—circulates regularly. The most reliable geolocations come from established news organizations and analyst networks that have invested in verification procedures and whose reputation depends on accuracy. When reading or sharing conflict footage, the location pin on a map is only as credible as the evidence behind it.