Supplementary material
======================
Description of files in this directory

Data Files
==============
dblp-2013-12-23.zip                                           - compressed dblp file from 23.12.2013 (original xml file: 1.4 GB)
dblp_id_gender_dict_v0.dict                                   - ID-gender python dictionary without cross checking heuristics
dblp_id_gender_dict_v1.dict                                   - ID-gender python dictionary with cross checking heuristics
name_gender_score.csv                                         - list of names, most likely gender and the score

Output
==============
dblp_id_name.csv                                              - list of IDs and the corresponding names
dblp_edges_maxAuthor20.csv                                    - edge list of co-author network with one entry for each collaboration and its year (multi graph of collaborations)
dblp_edges_maxAuthor20_simple.csv                             - edge list of co-author network with one entry for each collaboration and its year (simple graph of co-authorship)
dblp_edges_maxAuthor20_mentee_frequent.csv                    - edge list of student-mentor network (directed, first entry of each line is the ID of a student, followed by the mentor ID)
dblp_edges_maxAuthor20_mentee_frequent_known_degree_v1.csv    - degree statistics of student-mentor network (dictionary v1)
dblp_edges_maxAuthor20_mentee_frequent_known_degree_v0.csv    - degree statistics of student-mentor network (dictionary v0)
dblp_edges_maxAuthor20_simple_degree_1000_edges_v1.csv        - edge list of high degree known gender nodes in the dblp co-author graph (dictionary v1)
dblp_nodes_maxAuthor20_simple_degree_1000_nodes_v1.csv        - nodes list of high degree known gender nodes in the dblp co-author graph (dictionary v1)
dblp_edges_maxAuthor20_simple_degree_1000_edges_v0.csv        - edge list of high degree known gender nodes in the dblp co-author graph (dictionary v0)
dblp_nodes_maxAuthor20_simple_degree_1000_nodes_v0.csv        - nodes list of high degree known gender nodes in the dblp co-author graph  (dictionary v0)
degreesum_percentage_v1.csv                                   - data points of the figure depicting the degree sum percentages
FM_v1.csv                                                     - data points of the figure depicting the expected and observed number of mixed edges
tail_glass_ceiling_v1.csv                                     - data points of the figure depicting the percentage of women with a degree above k

Mathematica Code Files
====================
dblp_top_clean.nb                                             - generate networks for top nodes in dblp
dblp_bias_stat_clean.nb                                       - degree distribution in student-mentor graph
A_D_Proof.nb                                                  - C_B and C_R behavior 
Vcolor_simple_v1.m                                            - node colors in mathematica format   

Python Source
==============
parse_DBLP_xml.py                                             - parsing xml, creating collaboration graph for all publications with at most 20 co-authors
convert_to_simple_graph                                       - turn multi-graph (with possibly more than one edge per vertex pair) into a simple graph
create_mentor_file.py                                         - create the mentor graph based on the co-author graph and the first publication years
create_id_degree_file.py                                      - create a list of the degrees of each node
extract_top_graph.py                                          - compute the top nodes with a known gender and the induced co-author subgraph
compute_known_degree.py                                       - create the degree distribution file for degree/female/male/both/total of ids with known gender
README.txt                                                    - text file explaining how to run the python code.