Well, I'd start with two factors :-
- what will you use it for?
- slide or print?
If you're learning, and experimenting, try both approaches.
Firstly, be aware that slide can give fantastic results but tends to be less forgiving, exposure-wise. You can correct a lot with print, but slide is much less forgiving. You don't have anything like the latitude for incorrect exposure.
That's the downside. The upside is that there are some great films, but it helps to get the choice right, I love Velvia (mainly the original ISO 50) for scenic shots, but wouldn't use it for portraits or weddings. For that, personally, I prefer NPH400. It's good good tone balance especially for skin, is fast enough to give a lot of latitude for low-light use, such as wedding receptions, and as a print film, has wider latitude on the exposure than slide.
Another thing to think about is workflow. Are you staying non-digital for the whole process, up to an including final output, or are you going to scan, and Photoshop, and inkjet print? Because, there's any number of variants and none are right or wrong - simply whether they work for you and get the results you want.
For instance, a friend of mine shoots print film, has the film hand-processed and printed, scans the prints, tweaks in Photoshop and prints on wide-format Epson inkjets. And when I asked "why print photographically, scan and reprint on inkjets, because it seems a long way round" the answer, paraphrasing a bit, was "I get the results I want that way".
If logic was ... what works. works.
A bit more discussion revealed that my friend felt flatbed scanning a print was easier, more reliable and less fraught, and more cost-effective than scanning film provided you keep control over the print process. And the friend that does this is a successful enough photographer to not only have an international reputation, but to employ the lab techs to run their own lab. Funding a fully equipped darkroom and employing staff might go a bit beyond most of us, but doing your own colour developing isn't that hard, and mainly requires accurate temperature control and timings. You don't need that much in the way of equipment. Financially, it's probably worthwhile if you'll be doing it a lot, but chemicals have a limited longevity and you need to be doing enough to not waste chemicals.
So one option is .... develop your own film, use a cheap scanner to get a feel for which images are woreth extra investment, and get them hand printed to, say, 10x8. Then, scan them on a decent flatbed, Photoshop to tweak and then print for sale (or display, whatever).
Or, you can do what I've done in the past if l you can find the right local companies. A local print shop has a nice arrangement with a local pro. They do his hand-printing for him, as he does mainly slide, but he does the relatively small amount of slide processing they get for them. So I can get either slide or print film processed and, (if necessary) printed professionally, and locally. No mucking about with mail order.
Anyway, summing up, choose film and even film type based on the use you'll put it to, but consider different workflows if you do go digital. Providing the prints are done well, which generally means not mail-order, don't assume you necessarily need to scan the film itself. Scanning a good print works too. There's many ways to skin a cat.