Notebook version of demo_load_show_sheet.py

In [10]:

    

import pandas as pd

df = pd.read_excel("sheet_1_with_simple_logic.xls")
print(df)


    

    




   Feature1  Feature2 DecisionF1F2         TVShow  Decision2 Decision3  \
0       0.6       0.6         True      Hollyoaks          1      True   
1       0.4       0.6        False      hollyoaks          1     False   
2       0.3       0.4        False     Hollyoaks           0     False   
3       0.9       0.8         True      hollyoaks          1      True   
4       0.9       0.8         True     holly-oaks          0     False   
5       0.9       0.8         True  best TV shows          0     False   

                    Comment  
0                       NaN  
1                       NaN  
2  trailing-space on TVShow  
3                       NaN  
4        badly-spelt TVShow  
5                       NaN  

In [11]:

    

df.head() # this creates a Table view (non-interactive but prettier)
# NOTE! head shows 5 items by default and we have 6 items(!)


    

    
Out[11]:






  

    

      

      
Feature1
      
Feature2
      
DecisionF1F2
      
TVShow
      
Decision2
      
Decision3
      
Comment
    
  
  

    

      
0
      
0.6
      
0.6
      
True
      
Hollyoaks
      
1
      
True
      
NaN
    
    

      
1
      
0.4
      
0.6
      
False
      
hollyoaks
      
1
      
False
      
NaN
    
    

      
2
      
0.3
      
0.4
      
False
      
Hollyoaks
      
0
      
False
      
trailing-space on TVShow
    
    

      
3
      
0.9
      
0.8
      
True
      
hollyoaks
      
1
      
True
      
NaN
    
    

      
4
      
0.9
      
0.8
      
True
      
holly-oaks
      
0
      
False
      
badly-spelt TVShow
    
  

In [19]:

    

df.head(10)


    

    
Out[19]:






  

    

      

      
Feature1
      
Feature2
      
DecisionF1F2
      
TVShow
      
Decision2
      
Decision3
      
Comment
      
Feature1_Times_2
    
  
  

    

      
0
      
0.6
      
0.6
      
True
      
Hollyoaks
      
1
      
True
      
NaN
      
1.2
    
    

      
1
      
0.4
      
0.6
      
False
      
hollyoaks
      
1
      
False
      
NaN
      
0.8
    
    

      
2
      
0.3
      
0.4
      
False
      
Hollyoaks
      
0
      
False
      
trailing-space on TVShow
      
0.6
    
    

      
3
      
0.9
      
0.8
      
True
      
hollyoaks
      
1
      
True
      
NaN
      
1.8
    
    

      
4
      
0.9
      
0.8
      
True
      
holly-oaks
      
0
      
False
      
badly-spelt TVShow
      
1.8
    
    

      
5
      
0.9
      
0.8
      
True
      
best TV shows
      
0
      
False
      
NaN
      
1.8
    
  

In [12]:

    

print("Column names:", df.columns)


    

    




Column names: Index(['Feature1', 'Feature2', 'DecisionF1F2', 'TVShow', 'Decision2',
       'Decision3', 'Comment'],
      dtype='object')

In [13]:

    

print("Information about each row including data types:")
print("(note - type 'object' is catch-all that includes strings)")
df.info()


    

    




Information about each row including data types:
(note - type 'object' is catch-all that includes strings)
<class 'pandas.core.frame.DataFrame'>
Int64Index: 6 entries, 0 to 5
Data columns (total 7 columns):
Feature1        6 non-null float64
Feature2        6 non-null float64
DecisionF1F2    6 non-null bool
TVShow          6 non-null object
Decision2       6 non-null int64
Decision3       6 non-null bool
Comment         2 non-null object
dtypes: bool(2), float64(2), int64(1), object(2)
memory usage: 300.0+ bytes

In [14]:

    

print("\nWe can extract a column of data as a Series object:")
print(df['Feature1'])


    

    




We can extract a column of data as a Series object:
0    0.6
1    0.4
2    0.3
3    0.9
4    0.9
5    0.9
Name: Feature1, dtype: float64

In [15]:

    

row = df.ix[0]
print("\nWe can extract a row as a Python dictionary:")
print(row)


    

    




We can extract a row as a Python dictionary:
Feature1              0.6
Feature2              0.6
DecisionF1F2         True
TVShow          Hollyoaks
Decision2               1
Decision3            True
Comment               NaN
Name: 0, dtype: object

In [16]:

    

print("\nRow items, e.g. Feature1={feature1}".format(feature1=row['Feature1']))


    

    




Row items, e.g. Feature1=0.6000000000000001

In [17]:

    

def multiply_feature1_by_2(cell):
    return cell * 2

# we'll apply a function cell-by-cell to each cell in a Series (we pull out the Feature1 Series)
df['Feature1'].apply(multiply_feature1_by_2)
# note this doesn't change the DataFrame, it generates a new separate Series
# and here we just print it and then discard it


    

    
Out[17]:





0    1.2
1    0.8
2    0.6
3    1.8
4    1.8
5    1.8
Name: Feature1, dtype: float64

In [20]:

    

# we can assign the result back to the DataFrame as a new column
new_result = df['Feature1'].apply(multiply_feature1_by_2)
df['Feature1_Times_2'] = new_result
df.head(10)


    

    
Out[20]:






  

    

      

      
Feature1
      
Feature2
      
DecisionF1F2
      
TVShow
      
Decision2
      
Decision3
      
Comment
      
Feature1_Times_2
    
  
  

    

      
0
      
0.6
      
0.6
      
True
      
Hollyoaks
      
1
      
True
      
NaN
      
1.2
    
    

      
1
      
0.4
      
0.6
      
False
      
hollyoaks
      
1
      
False
      
NaN
      
0.8
    
    

      
2
      
0.3
      
0.4
      
False
      
Hollyoaks
      
0
      
False
      
trailing-space on TVShow
      
0.6
    
    

      
3
      
0.9
      
0.8
      
True
      
hollyoaks
      
1
      
True
      
NaN
      
1.8
    
    

      
4
      
0.9
      
0.8
      
True
      
holly-oaks
      
0
      
False
      
badly-spelt TVShow
      
1.8
    
    

      
5
      
0.9
      
0.8
      
True
      
best TV shows
      
0
      
False
      
NaN
      
1.8
    
  

In [ ]: