So I've been meaning to do a language family for a while and finally got around to sketching it out. Originally I wanted to do a proto-lang as it truly should be: constructed from the daughter languages. Well that proved unfeasible, so instead my current strategy is thinking of a word in a daughter language and then working backward. Not as cool, but a lot easier. I am going to try to keep the proto-lang fairly unspecified though. As far as world building is concerned, this family is the dominant language on the west continent, which goes from about the equator to quite south. There's other families there (and plenty of substrate influence) but it's still by far the largest. Anyway, some history.
Kikxotian explorers first encountered Cxernian speakers on the islands that stretch between the Moon People's home and the western continent. Eventually after reaching the main continent itself, some clever people started to notice minor similarities between the languages of the islands (the so called Yolean languages) and the languages the dominated the coast. More exploration eventually uncovered more language with similarities. After deciding that Proto-Viggo-Viggo wasn't a very informative name, they decided to name the family after the word for "person" in two of the languages, hence Yole from the language of Greater Rock Island (one of the Far Island languages and the first encountered by Kikxotian explorers) and Chort from Central Plateau Mountain Valley (the largest of the languages spoken in the mountains). Cxern (possibly [cχʷərn] but the actual pronunciation of X is unknown) is the reconstructed word for "person" in their reconstructed ancestor, hence the other name for the family, the Cxernian languages.
Proto-Yole-Chort was likely spoken some 6 to 7 thousand years before present in the mountains central to the continent. This placement is based on some of the reconstructed words, its location in the middle of the current spread, and because the greatest diversity is still found in that region. It's thought that the Cxernian languages spread with the invention of agriculture, especially the sweet potato. Some reconstructed words relating to this location and time period include *adiH "sweet potato" (seen in GRI as tete, in CPMV as adha and in Imperial Great Delta as arè), *QuSayaG "mountain peak" (GRI kwena "heaven", CPMV ahuśā "high holy place", IGD kohě "mountain") and *kelun "terrace, garden plot" (GRI hulo "earthworks", CPMV ūt "garden", IGD klung "paddy"). They likely had domesticated yaks or some other beast of burden before the family broke up based on words like *muNis "beast of burden" (reflexes include GRI mulisa "trade boat", CPMV amume "yak", IGD pùnê "water buffalo").
The internal classification of the Cxernian languages is not well understood. It's hypothesized that there's between 4 and 8 primary branches. The Badlands and Forest languages are well accepted as forming a Western Branch, and the Upriver and Delta languages are often (but controversially) linked together in an Eastern branch on basis of the treatment of *Q as /k/, and some lexical similarities (the Island family is occasionally placed here as well, and sometimes even as a branch of the Delta languages). The Mountain Valley languages are traditionally treated as their own branch. The so called Jungle Farmer and Jungle Hunter languages are often linked together but the evidence is lacking despite their close proximity (it's mainly typological as the so-called jungle languages are much more dependent marking than most Cxernian languages). Finally, the poorly attested Hill languages (known mostly as an ancient substrate of West Mountain Valley languages and as a small remnant dialect cluster deep in the mountains) are conventionally treated as part of the Western branch but it may actually be it's own primary branch or better treated as coordinate with Proto-Western instead of a daughter.
The subbranches of the primary branches are better understood. There's well over 30 surviving branches today, many of which have time depths of over 2000 years (and thus are comparable to Romance or Germanic languages in diversity). There's also many dead, but influential branches such as the so called "Old Raiders" language which left its loans and influence across the western part of the continent 2500 years ago (so before the Kikxotian branch of the Western Plains languages separated from its parent!) and most of the Hill branch. A lot of the subbranches seemed to have started diversifying around 2000 years ago, which incidentally is around when the Proto-Kikxotians headed east. Maybe there was some sort of global climate shock which influenced migratory patterns or something.
My current game plan is to define sound changes up to branching points, since I have their spatial and temporal positions well defined and from there create languages as needed. And if there's nowhere a language idea fits, I can just establish a new branch somewhere. Hard parts right now are grammar evolution and making sure I get both family internal and external loans right. There's a lot of other languages on the continent after all and I don't want to have a simple grammar in the protolanguage but I also want a very typologically diverse family, so we'll see what happens. Another issue is deciding what amount of sound change is sufficient for a new branch and over a given time period. By the end, I should have about Indo-European levels of diversity, so a lot. As for storage, I'll probably keep the proto-language's dictionary in a spreadsheet, since that will make tracking the different branches much easier (and in theory I can even automatically apply sound changes). But I think I have a good start.