In [1]:

    

%matplotlib inline
import matplotlib.pyplot as plt
plt.style.use('ggplot')
import warnings
warnings.filterwarnings("ignore")

%matplotlib inline
%load_ext autoreload
%autoreload 2

import csv, codecs
from matplotlib import colors, ticker, cm
import seaborn as sns
sns.set(color_codes=True)


    

In [2]:

    

# degree distribution plots
data_file = "dataset/node_stats.csv"

in_degrees = list()
out_degrees = list()
in_degrees_ref = list()
out_degrees_ref = list()
with codecs.open(data_file, "r", encoding="utf-8") as f:
    csv_reader = csv.reader(f,delimiter=';', quotechar='"')
    next(csv_reader, None)  # skip the headers
    for row in csv_reader:
        if len(row) < 10:
            continue
        internal_id,bid,title,author,lb,year,consultation,indegree,outdegree,degree = row
        in_degrees.append(int(indegree))
        out_degrees.append(int(outdegree))
        if lb == "True":
            in_degrees_ref.append(int(indegree))
            if int(outdegree)>0:
                out_degrees_ref.append(int(outdegree))


    

In [3]:

    

print(len(in_degrees))
print(sum(in_degrees))
print(sum(in_degrees_ref))


    

    




37626
71650
3933

In [4]:

    

sns.distplot(in_degrees, bins=150, kde=False, rug=False, hist=True, label="All")
sns.distplot(in_degrees_ref, bins=150, kde=False, rug=False, hist=True, color="red", label="Reference monographs")
plt.legend(loc="best",fontsize=14)
plt.ylabel("Number of cited monographs", fontsize=14)
plt.xlabel("In-degree", fontsize=14)
plt.xlim([0,100])
plt.tight_layout()


    

In [5]:

    

plt.hist(in_degrees, bins=150, log=True, label="All")
plt.hist(in_degrees_ref, bins=150, log=True, color="red", label="Reference monographs")
plt.legend(loc="best",fontsize=14)
plt.ylabel("Number of cited monographs", fontsize=14)
plt.xlabel("In-degree", fontsize=14)
plt.xlim([0,80])
plt.tight_layout()


    

In [6]:

    

sns.distplot(out_degrees_ref, bins=600, kde=False, rug=False, hist=True, label="Reference monographs")
plt.legend(loc="best",fontsize=14)
plt.ylabel("Number of citing monographs", fontsize=14)
plt.xlabel("Out-degree", fontsize=14)
plt.xlim([0,620])
plt.tight_layout()


    

In [ ]: