In [1]:

    

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline


    

In [2]:

    

df = pd.read_csv("./tendulkar_ODI.csv")
df.head()


    

    
Out[2]:







  

    

      

      
Unnamed: 0
      
Runs
      
Mins
      
BF
      
4s
      
6s
      
SR
      
Pos
      
Dismissal
      
Inns
      
Opposition
      
Ground
      
Start Date
    
  
  

    

      
0
      
1
      
0
      
-
      
2
      
0
      
0
      
0
      
5
      
caught
      
2
      
v Pakistan
      
Gujranwala
      
18-Dec-89
    
    

      
1
      
2
      
0
      
2
      
2
      
0
      
0
      
0
      
5
      
caught
      
2
      
v New Zealand
      
Dunedin
      
1-Mar-90
    
    

      
2
      
3
      
36
      
51
      
39
      
5
      
0
      
92.3
      
6
      
caught
      
1
      
v New Zealand
      
Wellington
      
6-Mar-90
    
    

      
3
      
4
      
19
      
38
      
35
      
1
      
1
      
54.28
      
4
      
bowled
      
2
      
v England
      
Leeds
      
18-Jul-90
    
    

      
4
      
5
      
31
      
31
      
26
      
3
      
0
      
119.23
      
6
      
bowled
      
2
      
v England
      
Nottingham
      
20-Jul-90
    
  

In [3]:

    

df["Runs"] = df["Runs"].str.replace("*", "").str.replace("T?DNB", "0").astype("int64")


    

In [4]:

    

df["4s"] = df["4s"].str.replace("-", "0").astype("int64")


    

In [5]:

    

plt.figure(figsize=(10,5))
sns.distplot(df["Runs"], bins=range(df["Runs"].min(), df["Runs"].max(), 10))
plt.xlabel("Runs Scored by Tendulkar")
plt.show()


    

In [6]:

    

plt.figure(figsize=(10,5))
sns.distplot(df["4s"], bins=range(df["4s"].min(), df["4s"].max()+1))
plt.xlabel("4s hit by Tendulkar")
plt.show()