2018 | Knight Foundation

How Disinformation Takes Flight on Twitter

Dataviz

Who’s behind the spread of “Fake News?” On the eve of the 2018 U.S. midterm election, the non-profit research group Knight Foundation analyzed Twitter data to investigate how disinformation spread during the 2016 presidential race. Their study—which included more than 10 million tweets from 700,000 Twitter accounts—found sobering statistics. With social media only continuing to grow in influence and specious activity showing no signs of slowing, Knight Foundation’s findings were urgent. They hired Accurat to help bring them to life for voters.

Brief

Knight Foundation’s study (conducted with the network analysis firm, Graphika) uncovered an army of hired–hands and bots responsible for the widespread propagation of false reports. The report linked 6.6 million tweets to disreputable sources in the month before the 2016 election, and an additional 4 million after, in a single month of 2017. With midterm elections looming, they were eager to share their findings. But couched in a 60–plus–page report, the average eligible voter was unlikely to stumble upon them. Knight Foundation needed a way to summarize their intel in a way that would engage the general public. They turned to Accurat to create a digital experience that would cut through the noise with powerful visuals.

Solution

Accurat’s designers and developers worked in close collaboration with Knight Foundation to adapt research into a user–friendly “scrollytelling” experience in just three months. To showcase the study’s discoveries, we summoned familiar iconography. The responsive microsite brought Knight Foundation’s report to life with visuals that called to mind birds in flight: a nod to the social media platform.

Focusing on the 24–hour period before the 2016 election, Accurat portrayed individual accounts as birds and their activity as flight paths. Tweets were shown as deviations in course: spikes in otherwise–straight lines. Comparing similar accounts, a user can discern nearly identical account movement. Bogus accounts acting in coordination thus resemble flocks of birds.

A second phase of the interactive experience shows the same user accounts’ activity over time, focusing on five busy days for relevant hashtags (i.e. #Wikileaks and #SethRich). Wrapping the same flight path visualization around a 24–hour clock graphic, users can see a clear difference in human versus bot activity. Outsized spikes appear unnatural both because of their size and their repetitiveness. While ordinary users tend to tweet sporadically during the day, bots and hired agents tweet at regular intervals, and oftentimes at odd hours of the night. We emphasized this detail with a gradient that darkened as nighttime hours ticked by.

To show the correlation among different interest groups—Trump supporters, hard-edge conservatives, white nationalist groups—Accurat color-coded accounts. Toward the end of the experience, we simplified the 44 hues that colored accounts with a blended image showing how disparate interests joined together to form a solid front.

Results

In adapting a lengthy, research report to a user-friendly, device-agnostic digital experience, Accurat helped further Knight Foundation’s mission of advancing journalism and preserving the integrity of communications in the 21st century. The report and Accurat’s take on it received coverage in Politco↗, NPR↗, and Adweek↗ NPR, and Adweek, among other outlets.

Website ↗

Team

Ruggero Castagnola

Sara Confalonieri

Gabriele Lippi

Giorgia Lupi

Alberto Massa

Tommaso Renzini

Gabriele Rossi

Cesare Soldini

Mariano Viola

Alessandro Zotta