In [2]:

    

%matplotlib inline


    

In [3]:

    

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.stats import chisquare
from scipy.stats import fisher_exact
import seaborn as sns
plt.rcParams['figure.figsize'] = 12, 7


    

In [4]:

    

pop = pd.read_csv('data/catalog.csv')
samp = pd.read_csv('data/clean_reception.csv')
len(pop)


    

    
Out[4]:





322

In [5]:

    

pop_auths = pop.groupby('Author').size()
samp_auths = samp.groupby('Author').size()
pop_prints = pop.groupby('Printer').size()
samp_prints = samp.groupby('Printer').size()
pop_place = pop.groupby('Place of Publication').size()
samp_place = samp.groupby('Place of Publication').size()


    

In [6]:

    

auths = pd.concat([samp_auths, pop_auths], axis=1, keys=["obsv", "exp"]).fillna(0)
top_auths = auths[auths.exp > 4]
bot_auths = auths[auths.exp < 5]
top_auths.ix["other"] = [sum(bot_auths.obsv.values), sum(bot_auths.exp.values)]
top_auths.exp = top_auths.exp.apply(lambda x: x / sum(top_auths.exp.values))
top_auths.obsv = top_auths.obsv.apply(lambda x: x / sum(top_auths.obsv.values))
top_auths, len(auths)
totals = auths.exp["Lope de Vega"] + auths.exp["Ledesma, Alonso de"] + auths.exp["Cervantes, Miguel"] + auths.exp["Aleman, Mateo"]
totals / 322.0, auths.exp["Lope de Vega Carpio, Fenix"]


    

    
Out[6]:





(0.63664596273291929, 100)

In [11]:

    

font = {'family' : 'normal',
        'weight' : 'bold',
        'size'   : 22}

plt.rc('font', **font)


    

In [19]:

    

sns.set_style("dark")
auths = top_auths.sort(columns="exp")
auths.plot(kind="bar")
plt.gcf().subplots_adjust(bottom=0.30)
plt.xticks(rotation='70')
plt.savefig("img/article/figure8_1.eps", format='eps', dpi=1000)


    

In [20]:

    

chisquare(auths.obsv.values, auths.exp.values)


    

    
Out[20]:





(0.036957990990821817, 0.99999999998231826)

In [21]:

    

prints = pd.concat([samp_prints, pop_prints], axis=1, keys=["obsv", "exp"]).fillna(0)
top_prints = prints[prints.exp > 4]
bot_prints = prints[prints.exp < 5]


    

In [22]:

    

top_prints.ix["other"] = [sum(bot_prints.obsv.values), sum(bot_prints.exp.values)]
top_prints.exp = top_prints.exp.apply(lambda x: x / sum(top_prints.exp.values))
top_prints.obsv = top_prints.obsv.apply(lambda x: x / sum(top_prints.obsv.values))
top_prints


    

    
Out[22]:






  

    

      

      
obsv
      
exp
    
  
  

    

      
Assiayn
      
 0.023810
      
 0.020761
    
    

      
Cormellas
      
 0.109524
      
 0.117647
    
    

      
Crasbeeck
      
 0.042857
      
 0.038062
    
    

      
Cuesta
      
 0.104762
      
 0.086505
    
    

      
ImprentaReal
      
 0.014286
      
 0.034602
    
    

      
Madrigal
      
 0.019048
      
 0.020761
    
    

      
Manescal
      
 0.033333
      
 0.044983
    
    

      
Martin
      
 0.100000
      
 0.083045
    
    

      
Mey
      
 0.019048
      
 0.038062
    
    

      
Rodriguez
      
 0.038095
      
 0.031142
    
    

      
Sanchez
      
 0.038095
      
 0.051903
    
    

      
Velpio
      
 0.042857
      
 0.034602
    
    

      
other
      
 0.414286
      
 0.397924
    
  

In [40]:

    

prints = top_prints.sort(columns="exp")
prints.plot(kind="bar")
plt.ylabel('% editions', size=18)
plt.xlabel('printer', size=18)
plt.gcf().subplots_adjust(bottom=0.15)
plt.xticks(rotation='35')
plt.ylim(0, 1)
plt.savefig("img/article/figure9.eps", format='eps', dpi=1000)


    

In [24]:

    

chisquare(prints.obsv.values, prints.exp.values)


    

    
Out[24]:





(0.041381279724033361, 0.99999999999989297)

In [23]:

    

place = pd.concat([samp_place, pop_place], axis=1, keys=["obsv", "exp"]).fillna(0)
top_place = place[place.exp > 4]
bot_place = place[place.exp < 5]
# totals = place.exp["Madrid"] + place.exp["Barcelona"] + place.exp["Lisboa"] + place.exp["Bruselas"] + place.exp["Valencia"] 
# totals / 322


    

    
Out[23]:





0.6428571428571429

In [26]:

    

# top_place.ix["other"] = [sum(bot_place.obsv.values), sum(bot_place.exp.values)]
# top_place.exp = top_place.exp.apply(lambda x: x / sum(top_place.exp.values))
# top_place.obsv = top_place.obsv.apply(lambda x: x / sum(top_place.obsv.values))
top_place.plot(kind="bar")


    

    
Out[26]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f95ec4d79b0>






    

In [27]:

    

place = top_place.sort(columns="exp")
place.plot(kind="bar")
plt.savefig("img/top_cities.png")


    

In [ ]:

    

chisquare(place.obsv.values, place.exp.values)


    

In [20]:

    

chisquare([100, 200], [200, 400])


    

    
Out[20]:





(150.0, 1.7336432457178413e-34)

In [ ]:

    




    

In [ ]: