Two years ago on a weekend, I built a tool to make it easier to evaluate Twitter accounts. Since then 36590 people used it to analyze 55390 different Twitter accounts.
Over the last months @mmkaradeniz and I made a new version. We launched it last night:
@accountanalysis doesn't use machine learning or AI „magic“. Instead of telling users if an account is authentic or not, it helps them to evaluate the accounts themselves.
It visualizes the different features (date, time, type, app, etc.) of Tweets to make them interpretable. /1
@accountanalysis The core concept is still the same, but it looks much better and is easier to operate. Not only for the users, but us developers as well. Enabling us to continuously roll out new features in the future.
Side-by-side screenshots of the old and new version. /2
@accountanalysis Some people had questions what different charts displayed and how they could be interpreted. There are now explanations for all charts, that can be toggled on and off at any time. /3
@accountanalysis The selected/retrieved Tweet count moved to the top to make it clear that not all Tweets of an account are analyzed. 3200 is the API limit by Twitter. It's possible to get more through the Premium API, but I don't think people would pay $100+ for the analysis of one account. /4