

Distinguish Fake vs. Real Audience
Sports & Entertainment
The online voting results for the annual NBA All-Star Game raised some questions around why certain unknown players received more votes than expected. Graphika helped the NBA determine the root cause and uncovered strengths and weaknesses behind certain players’ engagement strategy.
-
Challenge
Despite the NBA’s best efforts to filter out potential bot votes, concerns on the validity of vote results for a few lower profile athletes remained.
-
Program
Graphika’s algorithm was used to:
- Analyze more than 5 million tweets about the NBA All-Star voting;
- Discover hyperactive bot accounts which voted about 1,000 times per day;
- Identify voting entry points such as promotions by highly influential star Justin Bieber.
-
Outcome
As a result of Graphika’s analysis:
- The reasons for several players’ voting results were uncovered.
- Seemingly suspicious voting results for one player were proven to be credible while others were shown to be manipulated.
- The revealed vulnerabilities of the social media votes resulted in a change in the NBA’s voting system.
The Best of Graphika in Your Inbox.
Sign up for updates via our email newsletter