A cool data science project would be analyzing the dictionaries of all languages (at least those sharing an alphabet) and finding networks of cognates or shared

A cool data science project would be analyzing the dictionaries of all languages (at least those sharing an alphabet) and finding networks of cognates or shared words, i.e. building a connected graph where all vertices are words and edges represent a cognate between a word in two languages. I think the relevant keywords here would be "computational historical linguistics."