Link to github repo Know our data The design of the schema that contains the data from the FakeNews Corpus data set In our design of the schema, we have tried to reduce various anomalies as much as possible. We started by cramming all the data into a single relation, but we quickly found out that this created a lot of anomalies. Therefore, we iteratively decomposed this relations until we ended up with the design that can be viewed on figure \ref{fig:er-diagram-fakenews-corpus}.