based on initial evaluation I think there’s a few hundred tops that may be that way (if that)
Fascinating, I hadn't thought to look for duplicate IDs.
I'm not totally happy with my methodology either, but I'm happy to discuss it. And I agree that there isn't a perfect way to trim the data.
I think that is a valid way to analyze the data. However, I'm worried that when you start posting results, people are going to take it as characters meant for play whether that's what they are or not.
One thing I was having trouble with was the other class field returning not null when it appears to be null. Probably a special character present, but that may have more to do with my import than the data file itself.
I didn’t either. Stumbled upon it when I showed 3 level 1 characters were multiclassed. Turned out to be one listed 3 times and the class other field appeared blank.
I’ll have to think through some more on how I would do it.
Yea. But there’s no clear indicator which are and aren’t. We can try to eliminate some of the obvious ones not meant to be PCs. Though, there’s going to be some error bands no matter which way we go. Either you’re removing some valid characters or including some invalid ones. Or both.
One thing about multiclassing: I've seen multiple classes in the second class field. Like starting class is 'Fighter', and other class is 'Rogue/Wizard'. I see over 8k records like that in the full data.
Yea. I’m not seeing how to get the individual levels for those classes either?
Me neither. My guess is that 3+ classes is rare in the data, so the person who made the data set compressed the extra classes from the JSON into the other class field in the CSV.
What does the character ID signify? Could a player (or DM) use the same ID for different characters, the digital equivalent of erasing and reusing a character sheet? Or could a DM use a single ID for every character in a game so as to avoid the players having to pay for subscriptions? Etc.?
