Even only filtering by very concrete facts is going to not work, especially for the social sciences.
For example, the NISVS 2010 shows that quite factually, female rapes within the last year vastly outnumber male rapes... using a definition of rape that excludes the majority of cases where men are forced to have sex against their will. I highly doubt that a factual database would include such intricacies as the specific definitions used to give a fact.
Or for a less charged and more natural sciences inclined example, consider facts about gravity. Newtonian gravity isn't quite correct, but it can safely be treated as a fact by the general population who aren't familiar with the physics behind gravity.
For example, the NISVS 2010 shows that quite factually, female rapes within the last year vastly outnumber male rapes... using a definition of rape that excludes the majority of cases where men are forced to have sex against their will. I highly doubt that a factual database would include such intricacies as the specific definitions used to give a fact.
Or for a less charged and more natural sciences inclined example, consider facts about gravity. Newtonian gravity isn't quite correct, but it can safely be treated as a fact by the general population who aren't familiar with the physics behind gravity.