Codes to generate analysis and figures for Quantifying Institutional Gender Inequality in Contemporary Visual Art

Data

The data for reproduction can be found on Harvard Dataverse. After download the data, please put artists_gender_recog_xx.csv under the processed_data folder and the rest under the raw_data folder.

Main paper analysis and figures

To generate all results and figures of the main paper, run ./remove_birth.sh.

In this shell script, it contains the following steps:

Data processing

python 01-00-data_processing.py -t 0.6 -f: create merged dataframes (shows_with_person_with_gender, career_start_end, sales_with_person), filter data and print out basic information.

Select data

python 01-01-data_selection.py -t 0.6 -f -y 1990 -a: select data based on the career start year we select, create corresponding selected dataframe. Generate Figure 1.

Assign institutional gender inequality

python 03-00-gender_preference_assign.py -t 0.6 -f -y 1990: assign institution to gender inequality category

Analysis institutional gender inequality

python 03-01-gender_preference_stats.py -t 0.6 -f -y 1990 -p neutral -a and python 03-01-gender_preference_stats.py -t 0.6 -f -y 1990 -p balance -a: Basic statistics (count) of gender inequality category under gender-neutral and gender-balanced criteria. Generate Figure 2a, 2b 2c, Figure 3a, 3b.

For Figure 3a, 3b, further run

cd plot_country_preference
python plot_map_color.py

Artist co-ehixibion gender analysis

python 04-artist_career_gender_preference.py -t 0.6 -f -y 1990 -p neutral -n 10 -a: assign co-exhibition gender to artist, and further analysis. Generate Figure 5

Sales data preparation

python 05-00-sales_data_prep.py -t 0.6 -f -y 1990 -p neutral -n 10 -a: prepare sales data and output auction bias. Generate Figure 6a.

Logistic regression model

python 05-03-logit.py -t 0.6 -f -y 1990 -p neutral -n 10 -a: logistic regression model. Generate Table 2, Table 3 and Figure 6b, 6c.

Remaining figures in main paper

Figure 2d and Figure 3c

python 03-03-gender_preference_scatter_boundary.py 1990

Figure 4

python 06-network_viz.py 1990 generates the gml file of the network and we further process the visualization file.

Supplementary Material analysis and figures

Table 1

# process and get Men Artists/Women Artists/Ratio
python 01-00-data_processing.py --threshold 0.6 --remove_birth
python 01-00-data_processing.py --threshold 0.8 --remove_birth
python 01-00-data_processing.py --threshold 0.9 --remove_birth

# select, and get Selected Men Artists/Selected Women Artists/Ratio
python 01-01-data_selection.py -t 0.6 --remove_birth -y 1990
python 01-01-data_selection.py -t 0.8 --remove_birth -y 1990
python 01-01-data_selection.py -t 0.9 --remove_birth -y 1990

Results are printed in the command line.

Table 2

For Men Exhibitions/Women Exhibitions/Exhibition Ratio

python 01-01-data_selection.py -t 0.6 --remove_birth -y 1990
python 01-01-data_selection.py -t 0.8 --remove_birth -y 1990
python 01-01-data_selection.py -t 0.9 --remove_birth -y 1990

For Man-overrepresented Institutions/Woman-overrepresented Institutions/Gender-neutral Institutions

python 03-00-gender_preference_assign.py -t 0.9 --remove_birth

Results are printed in the command line.

Figure 1

Panel a, b, c

python generate_neighborhood_csv.py -f -p neutral
python generate_neighborhood_csv.py -f -p balance

Panel d

The codes for panel d is in SI_codes/MultiscaleMixing/Art Gender Network Multiscale.ipynb

Figure 2

Panel a, b, c

cd SI_codes
python 02-01-time_trend.py -f

Panel d, e

cd SI_codes
python career_start_year_stability.py

Figure 3

Run python 04-artist_career_gender_preference.py -t 0.6 --no-remove_birth -y 1990 -p balance -n 10 --no-save

Figure 4: solo exhibition

Run all scripts under SI_codes/solo_exhibitions

Table 3

python 05-03-logit.py -t 0.6 --no-remove_birth -y 1990 -p neutral -n 10 --no-save

Figure 5

Panel a, b

python 03-01-gender_preference_stats.py -t 0.6 -f -y 1990 -p neutral

Panel d, e

python 03-01-gender_preference_stats.py -t 0.6 -f -y 1990 -p balance

Panel c, f

cd SI_codes
python 03-01-gender_preference_expert_grade.py -p neutral
python 03-01-gender_preference_expert_grade.py -p balance

Figure 6

Panel a, b: python 04-artist_career_gender_preference.py --remove_birth -p neutral -n 15 Panel c, d: python 04-artist_career_gender_preference.py --remove_birth -p neutral -n 20

Figure 7

Run no_remove_birth.sh

License

Shield:

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.