In [3]:

    

%run "0. config.ipynb"


    

    




https://api.redmetrics.io/v1/event.csv?game=0fe53936-027f-461c-b087-b5b5737b2163&gameVersion=51b8a78a-8dd3-4a5e-9f41-01e6805e0f52&entityType=event
1866

In [101]:

    

players_stats = pd.DataFrame.from_csv("data/players_stats.csv", encoding="utf8")


    

In [102]:

    

players_stats.head()


    

    
Out[102]:






  

    

      

      
complete
      
configure
      
craft
      
death
      
equip
      
gotomooc
      
gotourl
      
pickup
      
reach
      
restart
      
selectmenu
      
start
      
switch
      
unequip
      
duration (seconds)
      
section
    
    

      
playerId
      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

    
  
  

    

      
0023dbb1-7f98-4cdb-8122-722f801f40b3
      
 0
      
 1
      
 0
      
  3
      
 0
      
 0
      
 0
      
 1
      
  2
      
 0
      
 1
      
 0
      
 0
      
 0
      
  175
      
 2
    
    

      
01b0c435-f0c0-4bfd-9189-86fc0d29b163
      
 0
      
 0
      
 0
      
  0
      
 0
      
 0
      
 0
      
 0
      
  0
      
 0
      
 0
      
 1
      
 0
      
 0
      
    0
      
 0
    
    

      
02bc076b-32aa-467a-bbc6-b746abedb7bd
      
 0
      
 0
      
 0
      
  0
      
 0
      
 0
      
 0
      
 0
      
  0
      
 0
      
 0
      
 1
      
 0
      
 0
      
    0
      
 0
    
    

      
02c6953a-0417-4858-8efb-1989be9f6b9d
      
 0
      
 1
      
 0
      
  0
      
 0
      
 0
      
 0
      
 0
      
  1
      
 0
      
 2
      
 1
      
 1
      
 0
      
   19
      
 1
    
    

      
0306db66-081d-4035-b30f-8358469d6ec3
      
 1
      
 0
      
 3
      
 17
      
 8
      
 0
      
 1
      
 5
      
 12
      
 1
      
 2
      
 0
      
 0
      
 2
      
 1030
      
 8
    
  

In [103]:

    

from sklearn.decomposition import PCA


    

In [104]:

    

players_stats.mean()


    

    
Out[104]:





complete                0.156028
configure               0.205674
craft                   0.312057
death                   4.198582
equip                   1.347518
gotomooc                0.007092
gotourl                 0.070922
pickup                  1.049645
reach                   3.425532
restart                 0.014184
selectmenu              1.184397
start                   0.680851
switch                  0.191489
unequip                 0.390071
duration (seconds)    570.815603
section                 2.773050
dtype: float64

In [94]:

    

X = players_stats.as_matrix()
n = X.shape[0]

pca = PCA(n_components=2)
X_r = pca.fit(X).transform(X)
X_vr = pca.explained_variance_ratio_

cmap = plt.get_cmap("YlOrRd")

print sum(X_vr)

X_std = 10 + 40 * players_stats["duration (seconds)"] / players_stats["duration (seconds)"].std()
X_color = cmap(players_stats["section"]/8)

plt.figure()
plt.scatter(X_r[:, 0], X_r[:, 1], s=X_std, c=X_color, alpha=.6, lw=0) #, c=c, label=target_name)
plt.title("all players (n = %s)" % (n))
plt.legend()


    

    




0.999987828303






    

In [96]:

    

for title in list_types:
  df = players_stats[ players_stats[title] > 0 ] 
  
  X = df.as_matrix()
  n = X.shape[0]
  
  # print X
  # print n
  
  if n > 1:
    pca = PCA(n_components=2)
    X_r = pca.fit(X).transform(X)

    # print sum(pca.explained_variance_ratio_)
    
    d = X_std[df.index]
    c = cmap(df["section"]/8)

    
    plt.figure()
    plt.scatter(X_r[:, 0], X_r[:, 1], s=d, c=c, alpha=.6 , lw=0) #, c=c, label=target_name)
    plt.title("%s > 0 (n = %s)" % (title, n))
    plt.legend()
  else:
    print "we skipped %s because there is no enough data" % title

plt.show();


    

    




we skipped gotomooc because there is no enough data






    













    













    













    













    













    













    













    













    













    













    













    













    

In [17]:

    

plt.hist(players_stats["reach"], bins=20)


    

    
Out[17]:





(array([ 64.,  27.,   3.,   8.,   9.,  22.,   5.,   1.,   0.,   0.,   0.,
          1.,   0.,   0.,   0.,   0.,   0.,   0.,   0.,   1.]),
 array([  0. ,   1.6,   3.2,   4.8,   6.4,   8. ,   9.6,  11.2,  12.8,
         14.4,  16. ,  17.6,  19.2,  20.8,  22.4,  24. ,  25.6,  27.2,
         28.8,  30.4,  32. ]),
 <a list of 20 Patch objects>)






    

In [92]:

    

for checkpoint in range(0,9):
  df = players_stats[ players_stats["section"] == checkpoint ] 
  
  X = df.as_matrix()
  n = X.shape[0]
  
  # print X
  # print n
  
  if n > 1:
    pca = PCA(n_components=2)
    X_r = pca.fit(X).transform(X)

    d = X_std[df.index]
    c = cmap(df["section"]/8)

    
    plt.figure()
    plt.scatter(X_r[:, 0], X_r[:, 1], s=d, alpha=.6 , lw=0) #, c=c, label=target_name)
    plt.title("checkpoint = %s (n = %s)" % (checkpoint, n))
    plt.legend()
  else:
    print "we skipped %s because there is no enough data" % checkpoint

plt.show();


    

    




we skipped 5 because there is no enough data