In [1]:

    

import pandas as pd
surveys_df = pd.read_csv("surveys.csv")


    

In [2]:

    

type(surveys_df)


    

    
Out[2]:





pandas.core.frame.DataFrame

In [3]:

    

surveys_df['sex'].dtype


    

    
Out[3]:





dtype('O')

In [4]:

    

surveys_df['record_id'].dtype


    

    
Out[4]:





dtype('int64')

In [5]:

    

surveys_df.dtypes


    

    
Out[5]:





record_id            int64
month                int64
day                  int64
year                 int64
plot_id              int64
species_id          object
sex                 object
hindfoot_length    float64
weight             float64
dtype: object

In [6]:

    

# convert the record_id field from an integer to a float
surveys_df['record_id'] = surveys_df['record_id'].astype('float64')
surveys_df['record_id'].dtype


    

    
Out[6]:





dtype('float64')

In [7]:

    

len(surveys_df[pd.isnull(surveys_df.weight)])


    

    
Out[7]:





3266

In [8]:

    

# how many rows have weight values?
len(surveys_df[surveys_df.weight> 0])


    

    
Out[8]:





32283

In [9]:

    

df1 = surveys_df.copy()
# fill all NaN values with 0
df1['weight'] = df1['weight'].fillna(0)


    

In [11]:

    

df1['weight'].mean()


    

    
Out[11]:





38.751976145601844

In [12]:

    

df1['weight'] = surveys_df['weight'].fillna(surveys_df['weight'].mean())


    

In [ ]: