/DataManagers

Fake news detection

Primary LanguageJupyter Notebook

DataManagers

Fake news detection

Fake news are one of the most discussed problems in modern society that doesn’t seem to have a solution. We tried to tackle this problem by developing a comprehensive framework which was trained on over 32.000 articles and 50.000 tweets collected by us. The functional algorithm was deployed into a webapp and a Twitter chatbot, applying multiple times the “Lean Startup Approach” to understand the best opportunities for this idea. The LSA methodology led us to interview brand managers of multinational corporations such as Coca-Cola and P&G, while also preparing a smoke test accessed by over 500 people. The final result was a concrete plan of action on how to create, manage and launch an automated fact-checking tool.