Both of these seem like a fantastic way to support your intrinsic biases.
That's a pretty grim way of looking at things. I could find a bigger problem with the fact that the main page has images of the candidates, of varying quality, aspect ratio, etc. with no attempt for normalization. The effect of candidate presentation on voter preference is a highly studied area in psychology and political science which should give us a lot more pause than the side effects of either of these tools.[1]
Both of these tools allow one to cluster and examine the data in a structured form. It doesn't prescribe or afford any type of interpretation. As both of us are heavily into Wikidata, how is this different than returning the value of a SPARQL query and relying on the user to be smart about using the output?
The alternative to having these sense-making tools is scrolling down a page with 70+ candidates trying to track five different parameters in one's brain, by relying only on memory. Or giving up and reverting to voting for those you recognize as friends. I don't think that's a good state of affairs.
-Andrew