In [1]:
%matplotlib inline
import csv
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
In [2]:
records = pd.read_csv('../data/fucking_final_dataset.csv')
In [3]:
records.head(1)
Out[3]:
In [4]:
plt.rcParams['figure.figsize'] = (12.0, 8.0)
In [5]:
records.groupby('canonical_country').count()['control_number'].sort_values(inplace=False, ascending=False).ix[:10].plot(kind="bar")
plt.savefig('images/fig1_overall_places.png')
In [9]:
records.groupby('slug').count()['control_number'].sort_values(inplace=False, ascending=False).ix[:25].plot(kind="bar")
plt.savefig('images/fig2_overall_cities.png')
In [6]:
records.sort_values('pub_year').groupby('pub_year').count()['control_number'].plot()
plt.savefig('images/fig3_pub_by_year.png')
In [8]:
records[records.pub_year < 1900].sort_values('pub_year').groupby('pub_year').count()['control_number'].plot()
Out[8]:
In [9]:
records[records.pub_year > 1900].sort_values('pub_year').groupby('pub_year').count()['control_number'].plot()
Out[9]:
In [12]:
top_slugs = records.groupby('slug').count()['control_number'].sort_values(inplace=False, ascending=False).ix[:10].index
In [13]:
top_producers = records[records.slug.isin(top_slugs)]
In [14]:
group_top_producers = top_producers.sort_values('pub_year').groupby(['slug', 'pub_year']).count()['control_number']
In [16]:
top_producer_df = pd.DataFrame({
'madrid,spain': group_top_producers.ix['madrid,spain'],
'barcelona,spain': group_top_producers.ix['barcelona,spain'],
'mexico,mexico': group_top_producers.ix['mexico,mexico'],
'buenos aires,argentina': group_top_producers.ix['buenos aires,argentina'],
'sao paulo,brazil': group_top_producers.ix['sao paulo,brazil'],
'santiago,chile': group_top_producers.ix['santiago,chile'],
'rio de janeiro,brazil': group_top_producers.ix['rio de janeiro,brazil'],
'bogota,colombia': group_top_producers.ix['bogota,colombia'],
'lisbon,portugal': group_top_producers.ix['lisbon,portugal'],
'valencia,spain': group_top_producers.ix['valencia,spain'],
}).fillna(0)
In [17]:
top_producer_df.plot()
plt.savefig('images/fig4_top_producers_by_year.png')
In [15]:
top_pre_1900_producers = top_producers[top_producers.pub_year < 1900]
group_top_pre_1900_producers = top_pre_1900_producers.sort_values('pub_year').groupby(['slug', 'pub_year']).count()['control_number']
top_pre_1900_producer_df = pd.DataFrame({
'madrid,spain': group_top_pre_1900_producers.ix['madrid,spain'],
'barcelona,spain': group_top_pre_1900_producers.ix['barcelona,spain'],
'mexico,mexico': group_top_pre_1900_producers.ix['mexico,mexico'],
'buenos aires,argentina': group_top_pre_1900_producers.ix['buenos aires,argentina'],
'sao paulo,brazil': group_top_pre_1900_producers.ix['sao paulo,brazil'],
'santiago,chile': group_top_pre_1900_producers.ix['santiago,chile'],
'rio de janeiro,brazil': group_top_pre_1900_producers.ix['rio de janeiro,brazil'],
'bogota,colombia': group_top_pre_1900_producers.ix['bogota,colombia'],
'lisbon,portugal': group_top_pre_1900_producers.ix['lisbon,portugal'],
'valencia,spain': group_top_pre_1900_producers.ix['valencia,spain'],
})
In [16]:
top_pre_1900_producer_df.plot()
Out[16]:
In [17]:
top_post_1900_producers = top_producers[top_producers.pub_year > 1900]
group_top_post_1900_producers = top_post_1900_producers.sort_values('pub_year').groupby(['slug', 'pub_year']).count()['control_number']
top_post_1900_producer_df = pd.DataFrame({
'madrid,spain': group_top_post_1900_producers.ix['madrid,spain'],
'barcelona,spain': group_top_post_1900_producers.ix['barcelona,spain'],
'mexico,mexico': group_top_post_1900_producers.ix['mexico,mexico'],
'buenos aires,argentina': group_top_post_1900_producers.ix['buenos aires,argentina'],
'sao paulo,brazil': group_top_post_1900_producers.ix['sao paulo,brazil'],
'santiago,chile': group_top_post_1900_producers.ix['santiago,chile'],
'rio de janeiro,brazil': group_top_post_1900_producers.ix['rio de janeiro,brazil'],
'bogota,colombia': group_top_post_1900_producers.ix['bogota,colombia'],
'lisbon,portugal': group_top_post_1900_producers.ix['lisbon,portugal'],
'valencia,spain': group_top_post_1900_producers.ix['valencia,spain'],
})
In [18]:
top_post_1900_producer_df.plot()
Out[18]:
In [21]:
counted_by_year = records.sort_values('pub_year').groupby('pub_year').count()['control_number']
len(counted_by_year.values)
Out[21]:
In [27]:
len(top_producer_df)
Out[27]:
In [29]:
top_producer_df
Out[29]:
In [30]:
# top_producer_df_percent = pd.DataFrame({
# 'madrid,spain': top_producer_df['madrid,spain'].divide(counted_by_year.values),
# 'barcelona,spain': top_producer_df['barcelona,spain'].divide(counted_by_year.values),
# 'mexico,mexico': top_producer_df['mexico,mexico'].divide(counted_by_year.values),
# 'buenos aires,argentina': top_producer_df['buenos aires,argentina'].divide(counted_by_year.values),
# 'sao paulo,brazil': top_producer_df['sao paulo,brazil'].divide(counted_by_year.values),
# 'santiago,chile': top_producer_df['santiago,chile'].divide(counted_by_year.values),
# 'rio de janeiro,brazil': top_producer_df['rio de janeiro,brazil'].divide(counted_by_year.values),
# 'bogota,colombia': top_producer_df['bogota,colombia'].divide(counted_by_year.values),
# 'lisbon,portugal': top_producer_df['lisbon,portugal'].divide(counted_by_year.values),
# 'valencia,spain': top_producer_df['valencia,spain'].divide(counted_by_year.values),
# })
In [ ]:
# top_producer_df_percent.plot.area(stacked=False)