For an investigation into misinformation on TikTok, the Documented team developed helpful methodologies for using AI to transcribe and analyze large amounts of downloaded videos.
Investigative journalists intending to cover social media and its societial effects must understand the intricacies of the companies that drive them, and think critically about novel angles of coverage.
Social media platforms are among the most difficult sites to scrape for data across the internet. A recent session at NICAR23 unveiled several dynamic new tools — including Junkipedia, a possible CrowdTangle replacement — that can perform a wealth of social media monitoring tasks, from tracking down who is behind harmful ads to identifying conspiracy groups or influencers spreading disinformation.
To investigate what the Russian invasion looked like to TikTok users in Russia and Ukraine, and how the content available differed from one side of the border to the other, a team of journalists from the Norwegian broadcasting company NRK set out to investigate the social networking site’s algorithms and how a user’s location provides differing digital narratives about the war.
Tracking the most popular data journalism stories on Twitter from September 6 to September 13, we found a visual project by the South China Morning Post explaining the global impact of the deadliest terrorist attack on American soil in US history, a Guardian investigation into the history of Canada’s residential schools, and a video project by Le Monde exploring the recent history of forest fires around the world.
Our NodeXL mapping from July 19 to 25, which tracks the most popular data journalism stories on Twitter each week, found imaginative data-driven coverage of the Olympics by The Washington Post, The New York Times, and other major outlets. In this edition, we also feature a piece by Reuters Graphics on how a Chinese lake has been decimated by sand mining, a cross-border investigation into the billion Euro business of elder care, and a deep dive into TikTok’s secretive algorithm by The Wall Street Journal.
TikTok, a video-sharing site where users can post videos of themselves dancing, lip syncing, and doing viral challenges, has seen a surge in popularity in recent years. While many posts are focused on jokes and music, TikTok has surpassed 2 billion downloads and is popular around the world, which presents opportunities for the open source research community to use the platform in investigations. This guide explains how.
In a GIJN masterclass webinar presented to 700 journalists from 94 countries, BBC internet research expert Paul Myers described the open source tools, syntax tricks, and search techniques that can help unearth elusive content, images, and social media posts. Myers cited searches for information on COVID-19 in numerous examples, and stressed the need for logical thought in order to unlock the power of advanced search tools.
The popularity of TikTok has surged during the pandemic, and one particular “data investigation” clip has gone viral. Our NodeXL #ddj mapping from April 20 to 26 found TikTok user Rebecca fact-checking a woman’s claim about the COVID-19 quarantine and her grey hair roots, the Coronavirus Fact-Checking Alliance visualizing the thousands of fact checks they have produced during the pandemic, The New York Times analyzing United States President Donald Trump’s messages about the country’s coronavirus response, and FiveThirtyEight examining how concerned Americans are about the coronavirus compared to the economy.
