From day one, civilians and members of the military have been recording the aggressive Russian invasion in Ukraine. Social media has been flooded with videos from both sides.

Keeping track of all the movements and separating what is real from misinformation is an enormous challenge. We know from experience that our users struggle to keep up with complex cases that extend over a long period of time.

In February, we decided to develop an interactive map to present the dramatic developments. On a map of Ukraine, we showcase videos we have successfully verified, meaning we can put a time and place on it.

Verifying content

The amount of false war content emerging is huge, and we are all possible victims of disinformation. As publishers, our responsibility is to share the facts and distinguish this from false information.

Without reporters on the ground, we cannot know for sure what is going on. Verifying emerging content is becoming more important and difficult than ever before.

“Over the past years, we have experienced a significant increase in fake images from different news events in circulation on social media,” said Thomas Andreassen, who has been a picture editor at VG for 15 years.

Forging imagery has never been easier. Many apps and programmes exist that can easily fake the reality of what is portrayed in photos. As a result, we have increased our capacity on how to identify fake images.

“Currently, we are putting in a lot of resources to verify images and videos related to the ongoing war in Ukraine,” Andreassen said.

“Visual forensics”

Like many publishers, VG put together a task force with journalists working solely on verifying content. Over the past few weeks, video journalist Stig Øystein Schmidt has engaged in “visual forensics,” analysing and cross-referencing a large number of videos.

“Video verification has always been a part of our daily work routine. But the war in Ukraine has raised the bar for us and made us become even more systematic and thorough in our verification work,” Schmidt said. “We have made a small task force so we are able to work with the massive amount of videos that are flowing from the war zone.”

Schmidt and the team are using the open-source intelligence (OSINT) method. This means using tools that are available to the public and useful for supporting intelligent analysis.

The “forensic” approach

A journalistic and “forensic” approach allows us to pinpoint the location of videos emerging from the war and verify their authenticity. The most commonly used tools include:

• Metadata.
• Observations.
• Traces of manipulation.
• Weather.
• Geographic information.
• Open-source information.
• Web sites.
• Social media like the Snap Map, Instagram Stories, TikTok, Twitter, and Telegram.
• Search engines, especially Google Earth and Street View.
• Reverse image search, Google Lens, and TinEye.

The goal is to answer the important journalistic questions: Who? What? Where? When?

An example of geo-verifying

One of the videos Schmidt has been working on is the story of Russian soldiers allegedly holding hand grenades over their heads and the Konotop mayor giving the ultimatum: “If you show resistance, we will destroy the city.”

This is the process Schimdt followed to verify the videos and find its geolocation:

1. Notice buildings in background.

2. Notice the blue sign with letters on the wall.

3. Get help from a translator to understand what the sign means.

4. Find the Cyrillic alphabet on Wikipedia.

5. Get help from a translator to copy the correct Cyrillic letters.

6. Place the correct Cyrillic letters in Google maps and add “Konotop” to narrow the search.

7. Use Google Street View to find the correct building with the sign.

8. Turn around within Google Street View to see the other building across the street.

9. In videos, check to see if the weather is the same.
10. In videos, check to see if the mayor’s outfit matches.

Developing the map

When our journalists have been through a verification process similar to the one described above and we know the video’s geolocation, we pin it on our interactive map of Ukraine. The map was developed at the end of February by a cross-functional team of editorial developers.

We use a simple solution as our administrative tool for the map: a Google spreadsheet. Data from the sheet — such as title, geolocation, and video ID — are automatically fetched in JSON format, which makes it easy for the code to loop through the video list and place the items on the map.

The solution is coded in JavaScript on top of open-source frameworks such as React, Next.js, and MapLibre GL. The underlying map is based on data from MapTiler and the OpenStreetMap project, while colouring of occupied areas was created in QGIS, an open-source geo-data tool.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .