medialab / gender_art Goto Github PK

• calculate gender ratios on the total database
• same ratio though time by the acquisition date : how many artworks from women had been acquired by year / artworks from men had been acquired by year
• same ratio by institutions (through time)
• same ratio by artworktypes (through time)

Y a xis is number of artworks
X axis is dif between acquisition and creation date
Do one barchart by gender plus one for weird cases (multi gender or groups...)

Python3 script which does :

• retireve personnal account info from config
• generate a token and store into a file see the documentation
• craft the url with pagination
• set up a cache system to avoid restarting from scracth if we interupt the script
• set up a retry mechanism (see the bash script)
• store the data into json files as the bash script we doing it

• compute an average date for each artists which is the avergae of birthdate, deathdate, creation dates of all artworks
• depict a barchart of number of artist by average year (distribution control)
• for each important data field depict a barchart which represents the number of artists which does not have this data field available by the average year

• do we have group compisition
• if yes, how to count them gender wise ? consider a group as female if at least is present or majority or only groups which are composed of female artists ?

• API connection settings
• Reuse the existing collection script

response['totalCount'] equals to 379366
sum of response['aggregations']['gender'] equals to 364829

unique_artworks.py excludes series of artworks when creating dataset.
code in lines 37-42:
separable_artworks_groups = db.Artwork.aggregate([
{"$match": {"type":'separable'}},
{'$group': dict( [('_id','ensemble_id')] + [(k,{'$first':'$%s'%k}) for k,v in project.items()])},
{"$project":project}
])
should create a list of unique artworks from series, but it takes only the first one individual artworks of first series. It also creates '_id': 'ensemble_id' instead of the numerical id of series.

As a result all other series of artworks are excluded from the dataset.