vmfelso
commented
2 years ago Hey Jonas, a lot of separate stuff in this issue, I'll try not to miss anything ;) The good news is the data scheme was just written down by me by hand so nothing big to change. I'll try to list things in order and by subject now:
Typos, German mistakes etc: 1, 2, 4a, 6, 9b: sounds good, thank you -- I've fixed them in the README.md and jinja template [DONE]
Address questions: 3, 10: We can separate these out, I think we're getting them from a different data source -> separate issue
Rest of 4: Family membership: in the app we do have a preset category, although they are gendered I think it is fine to prefill this with the person but not the relationship if we don't know the gender. The way we are uniquely idenifying people is a combination of url + first name + last name. I think having a way to select "Who" is the relation would be nice and will look into this -> separate issue
5: can add this, I believe we will keep it binary because of historical reasons and that they are mostly possible to identify [DONE, I wasn't sure about how to word this in German but gave a try.]
7: I don't really have enough context of what other academic titles (MSc, Diplom Ing?) there are, but if I remember correctly there was a nun (or other religious person) in the database with an extra title so maybe we could be more inclusive about that. What do you think @JStet ? [-> let's keep discussion here]
8: Yes, this is such a big one to miss! Thank you for noticing it. -> separate issue
9a: I guess the idea is we will know from our document, but I'm not sure how far we will get with that. What do you think @JStet ? [-> let's keep discussion here]
jstet
"geburtsjahr" shows up twice
"geburtsmonat" instead of "geburtsmonate" would be grammatically correct
address a. we could separate the address into street and street number? b. we should highlight that the street names should not be abbreviated (eg. Seestraße instead of Seestr.)
Family: a. "Verwandschaftsbeziehung" or "verschwandschaftsart" instead of "Verwanschaftsgrad" b. Use predefined relationship types (have you maybe already done that?) and add them to the scheme c. We could use gender neutral names for relationship types, e.g. EhepartnerIn instead of separating between Ehemann und Ehefrau or Geschwister instead of Bruder and Schwester d. maybe we could add an id for every victim and then use it to reference to victims in the family attribute. the url is a nice unique identifier for example (edit: nvm, there are urls for families)
add gender a) did we decide on using binary categories?
write the variables with the first character in uppercase?
academic title instead of doctor title? (although there are only doctor titles)
Did we forget the stages of flight? See this document
Stations: a) add type of station? b) "Ort" instead of "haftort", because more general.
Adresse: a) add PLZ?
Sry I'm adding stuff while cleaning and transforming the scraped data, this list escalated quickly :S