You can either decide by what is currently in demand in the industry and then pick a project that you can exercise that language with or you can think of a project you’d like to do and then go by what the best language is for a given project.
In the end, languages are just like different wrenches. First you have to learn how to use a wrench, size or features don’t matter much at this point (unless you already know that you want to become an expert with one particular wrench).
I think starting a new project is way easier than contributing to an existing one.
Definitive roadmap (for the lazy people edition):