katero
Jun 30, 2026

Inside the Damascus Dossier: From leaked images to verified data - ICIJ

Data methodology

Inside the Damascus Dossier: From leaked images to verified data

ICIJ and its partners organized and analyzed thousands of chilling photographs to assemble comprehensive victim lists and quantify the human toll behind a sensitive data leak.

By and Image: Molly Crabapple / ICIJ December 10, 2025

One year after the fall of former Syrian president Bashar Assad, a new leak shines a light on the brutality of Assad’s rule and the lengths his regime took to catalog the bodies of those killed during his last decade in power.

The Damascus Dossier is an investigation based on a cache of more than 134,000 records obtained by German broadcaster NDR and shared with the International Consortium of Investigative Journalists and 24 media partners. The leak contains the largest photographic archive of bodies of Syrian detainees who died between 2015 and December 2024 — more than 10,200 bodies in 33,000 photographs. This is how ICIJ and its partners arrived at those figures.

The Syrian civil war began in 2011 following the Arab Spring protests and ended on Dec. 8, 2024, when the Hay’at Tahrir al-Sham militia toppled Assad’s government, leaving Assad and his top officials to flee to asylum in Russia. At least 160,000 Syrians were arrested and disappeared during those 13 years.

The chilling archive of photos, taken inside Assad’s system of incarceration, includes images of thousands of bodies, nearly all men and what appear to be a few teenage boys. A cross-border team of reporters and editors reviewed hundreds of these images and found that the bodies were mostly naked, emaciated and appeared to have been abused, lying on the floor, and sometimes covered in flies or fly droppings. Some appeared to have been starved to death; others likely died under torture. All died at the hands of Assad and his functionaries. The dead included at least one newborn.

Stripped of their dignity even in death, prisoners were also stripped of their names, reduced to a detainee number visible on white labels usually placed on their chests or foreheads.

Counting the dead

As part of the investigation, members of the ICIJ technology and data teams received a dataset of folders containing photos. To assess how many people were among those photographed, the data team analyzed the folder structure, as each folder was organized by year, then month, then day and photographer. Each directory within the dataset shared with ICIJ was named with Arabic characters and numbers representing the detainee numbers assigned to each victim. Some folders were named with a single detainee number; some with several detainee numbers, and some had more than a dozen detainee numbers in a single folder name.

ICIJ’s data team mapped out the file paths from the dataset. Reporters separated out the year folders, then the months, then the folders containing the detainee photos. They then created a list of the long Arabic folder names and put them into a spreadsheet. Once the team translated the folder names to English, they included all those labeled “detainees,” or “detainee,” eventually arriving at the figure of more than 33,000 detainee photographs.

Photo of a desolate, rocky site with three people walking in the background, and remnants of clothes visible in the forgeround.
A mass grave outside of the town of al-Otaiba, east of Damascus. The victims were killed in a February 2014 ambush by Syrian military forces. Image: Aref Tammawi / ICIJ

The data team then analyzed the photos and manually counted the detainee numbers in each folder name. That process allowed us to determine that the photographs in the leak contained more than 10,200 bodies, 70% of which were dated from 2015 to 2017.

This dataset was different from any other ICIJ has received in its 25-year history. The challenges were emotional as well as intellectual.

Such work has a cumulative effect. One member of the data team, having seen thousands, reached a point at which she could not bear to look at any more.

The images “burn themselves into your mind,” said Benedikt Strunz, an investigative reporter and editor at NDR. “Because you see things in them that shouldn’t really exist.”

Analyzing the contents of the images

Twelve journalists from ICIJ, NDR and Süddeutsche Zeitung agreed to examine and analyze a random sample of the photographs of the bodies to provide deeper insight into the trove of images. Prior to the analysis, the journalists participated in an online training session to prepare themselves to report on potentially traumatic content.

To have a representative statistical sample, ICIJ’s data team selected 540 of the 33,000 photographs using the RAND Excel function created for that purpose. This sample size allows for a 98% confidence level that what the journalists saw in the sample set could be fairly used to describe the photos in the entire dataset.

Other posts