In [18]:

    

# ***Chris Barber concieved of and carried out this analysis first (I am replicating it here and any bugs are my own)***

#Data sources:
#http://www.erswhitebook.org/files/public/Supplementary%20Material/Interstital%20Lung%20Disease/WB_ILD_supplement.xlsx
#http://www.erswhitebook.org/files/public/Supplementary%20Material/Occupational%20Lung%20Disease/WB_occupation_supplement.xlsx
    
import pandas as pd
%matplotlib inline
from scipy.stats.stats import pearsonr

meso = pd.read_csv('../data/ers_whitebook/eu_meso_mort.csv', skiprows=2)
ipf = pd.read_csv('../data/ers_whitebook/eu_ipf_mort.csv', skiprows=2)

meso = meso[['Country', 'Mesothelioma']][:28] #eu per 100,000 age standardised meso mort by country
ipf = ipf[['Country', 'Idiopathic fibrosing alveolitis and other ILD']][:28] #eu per 100,000 age standardised meso mort by country

ipfmeso = pd.merge(ipf,meso).dropna() #this drops Greece because don't have IPF and Meso data for Greece

ipfmeso.columns = ['Country','IPF','Meso']


    

In [19]:

    

ipfmeso.head()


    

    
Out[19]:






  

    

      

      
Country
      
IPF
      
Meso
    
  
  

    

      
0
      
Austria
      
0.84
      
0.80
    
    

      
1
      
Belgium
      
1.59
      
1.26
    
    

      
2
      
Bulgaria
      
0.12
      
0.06
    
    

      
3
      
Cyprus
      
2.57
      
0.63
    
    

      
4
      
Czech Republic
      
1.27
      
0.29
    
  

In [20]:

    

ipfmeso.Country.nunique()


    

    
Out[20]:





27

In [21]:

    

ipfmeso.plot(kind='scatter', x='Meso', y='IPF')


    

    
Out[21]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f820ab23b10>






    

In [22]:

    

ipfmeso.corr(method='pearson')


    

    
Out[22]:






  

    

      

      
IPF
      
Meso
    
  
  

    

      
IPF
      
1.000000
      
0.508095
    
    

      
Meso
      
0.508095
      
1.000000
    
  

In [23]:

    

pearsonr(ipfmeso.Meso, ipfmeso.IPF)


    

    
Out[23]:





(0.50809485103217877, 0.0068135519757501024)