The project uses public data from the FEC, and the code is all open source. I'd invite you to verify for yourself that I am not cherrypicking data.
You also completely failed to address my larger point, which is that the data or code is not flawed, your assumptions are.