Hello everyone,

September has begun, which means a new batch of challenges is ready for you to try. This month, we’re taking a new angle — one that’s very high up.

Also in this newsletter: the solutions to last month’s exercises. As always, we explain the methods used to solve them, drawing on the solutions you shared with us online.

Bellingcat is the only investigative organization where the collective builds society's capacity to find the truth. We turn obsessive curiosity into a teachable research skill, build the public community space to practice it, and continually innovate solutions to drive discovery forward. You can support our work here: https://www.bellingcat.com/donate/

This month, we’re exploring one of the most widely used open source research methods: satellite imagery analysis. It can be a valuable tool in conflict analysis, like in our research into the ongoing demolitions in southern Lebanon. But you can also use it to track ships, analyse earthquake damage and identify landslides. 

Jake Godin is part of Bellingcat's Conflict and Human Rights Team. Much of his investigative work involves satellite imagery, and he regularly monitors recent imagery of conflict-affected areas for significant changes. He created five challenges to explore some of the tools and techniques he uses in his investigations.

Bellingcat’s toolkit has some helpful information about satellite imagery providers. Check it out here: https://bellingcat.gitbook.io/toolkit/categories/maps-and-satellites/satellite-imagery 

Good luck! Challenge.bellingcat.com  

In our Discord server, you’ll find others working on the challenges. This is where you can ask for help and share your findings. 

SPOILER ALERT: From this point on, we’ll be discussing last month’s challenges. If you still plan to complete them, you may want to skip this section.

Here are the answers to the “Editing Errors” challenges:

For ‘The Road Ahead’, Faaiz K. explains how he used Google Reverse Image Search to identify the general area of Cape Town. But to find the exact location he needed to take a more manual approach. He matched the mountain featured in the photo to a location using Google Street View. The exact spot was on Chapman’s Peak Drive, and following this road south revealed a tollgate blocking the path. 

In ‘Terrain Tactics’, Faaiz K. noted all the visual clues in the image. He highlighted the mountains in the background, what appeared to be a salt basin, a hiking trail and branches. Faaiz entered this list of features into Gemini, which generated a list of places where these characteristics are present. One of the locations seemed like a good fit, with a salt basin and a mountain range nearby. Using photos posted to Google Maps of a popular tourist destination in the area, he managed to identify the exact location.

For ‘Peak Performance’, the description hints that the location is near Triglav, the highest mountain in Slovenia. Jurica Radović describes how he found the answer using Google Reverse Image Search. Google suggested a blog that included pictures of hikes in the region, as well as a Wikiloc trail. Following this trail on Google Maps, he found a Google Street View photosphere from the area, which he used to compare the visible mountain ridge with the peak in the challenge image.

‘False Front’ is a photo taken straight from a Bellingcat investigation. The investigation narrows down the general area, but to pinpoint the exact coordinates Zetroc Leirbag had to take a different approach. Checking the metadata proved to be a dead end, so he turned to the vegetation. The plants in the background were identified using Google Lens and a plant identification website. The two tools suggested slightly different species, but both are mainly found in Florida, according to iNaturalist. Combined with the Bellingcat investigation, this gave a good indication of the region to search.

Zetroc then used Overpass Turbo, a tool for querying OpenStreetMap data. He wrote a query with the help from an AI assistant that identified 24 bridges with matching features in Florida and found one good match. Although the number painted on the bridge was different, Google Street View shows the bridge that matches the photo just down the road. 

For ‘Not in Kansas’, we turn to DamCoffee’s Medium write-up. They thought the challenge photo must have been taken in a desert, but a reverse image search suggested Mars. Shifting their attention to outer space, DamCoffee used an interactive map from the European Space Agency to explore the Jezero Crater. This search produced no matches, so they turned to the metadata and found the original filename. The filename included the day the photo was taken, and that it was captured by Curiosity’s Mastcam Right camera. Using NASA’s Curiosity rover map, they found that the rover was near the prominent feature Mount Remarkable on the relevant day.

That’s it for this month’s Bellingcat Challenge Newsletter. We’d love to hear your feedback on the challenges. You can also join us on Discord and let us know if you have ideas for future challenges.

If someone forwarded you this newsletter and you’d like to subscribe, click here.

Before we go, here are some links to other current projects from Bellingcat: