[ Python 常見問題 ] Pandas - How to collect frequency from two columns of a dataframe

 Question

Consider we have below Dataframe:


Two questions here:
* Question1

How to create another column hp_rank to map column horsepower to High (150+), Medium (100-150) and Low (0-100)

* Question2

How to count frequency of the combination of column hp_rank and drive-wheels to show below table

HowTo
For Question1, the intuitive approach will look like:

view plaincopy to clipboardprint?

1. from collections import defaultdict  
2.   
3. data['hp_rank'] = data.apply(  
4.     lambda r: 'Low (0-100)' if r.horsepower <= 100 else 'Medium (100-150)' if r.horsepower <= 150 else 'High (150+)',  
5.     axis=1  
6. )  

A better approach is by using API pandas.cut:

view plaincopy to clipboardprint?

1. bin_labels = ['Low (0-100)', 'Medium (100-150)', 'High (150+)']  
2. data['hp_rank'] = pd.cut(  
3.     data['horsepower'],  
4.     bins=[0, 100, 150, data['horsepower'].max()],  
5.     labels=bin_labels  
6. )  

Regarding the Question2, you can leverage another API pandas.pivot_table:

view plaincopy to clipboardprint?

1. data['value'] = 1  
2. pd.pivot_table(data, values='value', columns=['drive-wheels'], index='hp_rank', fill_value=0, aggfunc=np.sum)