In [1]:

    

%run "0. config.ipynb"


    

    




1887

In [2]:

    

players_stats = pd.DataFrame.from_csv("data/players_stats.csv")


    

In [4]:

    

print len(players_stats["sessionId"].unique())


    

    




53

In [57]:

    

players_stats["count"] = 1

session_id = players_stats.groupby(["sessionId"]).agg({ "count": np.sum, "complete": np.sum })

session_id.head()


    

    
Out[57]:






  

    

      

      
count
      
complete
    
    

      
sessionId
      

      

    
  
  

    

      
05ba6db4-ac71-4953-a524-a4a9ebd9df91
      
 1
      
 0
    
    

      
069e1b60-538e-41d3-a6f5-29becc606521
      
 1
      
 0
    
    

      
0baf8bd8-8697-4628-99cb-538f8f5c0f24
      
 2
      
 0
    
    

      
0e1f8f26-dfd2-4265-b916-317435bf50ab
      
 2
      
 1
    
    

      
0ffa7953-c993-4a73-a93e-b3917484fdfb
      
 1
      
 0
    
  

In [58]:

    

session_id.hist("complete")


    

    
Out[58]:





array([[<matplotlib.axes._subplots.AxesSubplot object at 0x1134ea6d0>]], dtype=object)






    

In [60]:

    

session_id.plot(kind="scatter", x="count", y="complete")


    

    
Out[60]:





<matplotlib.axes._subplots.AxesSubplot at 0x1135a1510>






    

In [56]:

    

df = pd.DataFrame()

df["class"] = session_id["count"]

df = df.reset_index()
df = df.drop("sessionId",1)
df["count"] = 1

df = df.groupby(["class"]).agg({ "count": np.sum })

print np.sum([ i * x["count"] for i,x in df.iterrows() ])

A = df.reset_index().as_matrix()
#print A
print A[:,0].dot(A[:,1].T)

df.head(20)


    

    




97
97






    
Out[56]:






  

    

      

      
count
    
    

      
class
      

    
  
  

    

      
1 
      
 39
    
    

      
2 
      
  9
    
    

      
4 
      
  1
    
    

      
9 
      
  1
    
    

      
13
      
  1
    
    

      
14
      
  1