My understanding is that TIOBE is primarily based on number of search engine results. Other results like number of active (a commit within the last 6-12 months) open source repos or number of new repos in a given language are other metrics worth considering. The project you linked seems to be one like this.

A high search rate doesn't necessarily mean high rate of real world usage. Correlating across multiple metrics would be a better way to measure popularity.

Although popularity itself may still be a weak signal depending on your purposes.