Ex - GroupBy¶

Check out Alcohol Consumption Exercises Video Tutorial to watch a data scientist go through the exercises

Introduction:¶

GroupBy can be summarized as Split-Apply-Combine.

Special thanks to: https://github.com/justmarkham for sharing the dataset and materials.

Check out this Diagram

Step 1. Import the necessary libraries¶

In [1]:

import pandas as pd

Step 2. Import the dataset from this address.¶

Step 3. Assign it to a variable called drinks.¶

In [4]:

drinks = pd.read_csv('https://raw.githubusercontent.com/justmarkham/DAT8/master/data/drinks.csv')
drinks.head()

Out[4]:

	country	beer_servings	spirit_servings	wine_servings	total_litres_of_pure_alcohol	continent
0	Afghanistan	0	0	0	0.0	AS
1	Albania	89	132	54	4.9	EU
2	Algeria	25	0	14	0.7	AF
3	Andorra	245	138	312	12.4	EU
4	Angola	217	57	45	5.9	AF

Step 4. Which continent drinks more beer on average?¶

In [6]:

drinks.groupby('continent').beer_servings.mean()

Out[6]:

continent
AF     61.471698
AS     37.045455
EU    193.777778
OC     89.687500
SA    175.083333
Name: beer_servings, dtype: float64

Step 5. For each continent print the statistics for wine consumption.¶

In [9]:

drinks.groupby('continent').wine_servings.describe()

Out[9]:

continent       
AF         count     53.000000
           mean      16.264151
           std       38.846419
           min        0.000000
           25%        1.000000
           50%        2.000000
           75%       13.000000
           max      233.000000
AS         count     44.000000
           mean       9.068182
           std       21.667034
           min        0.000000
           25%        0.000000
           50%        1.000000
           75%        8.000000
           max      123.000000
EU         count     45.000000
           mean     142.222222
           std       97.421738
           min        0.000000
           25%       59.000000
           50%      128.000000
           75%      195.000000
           max      370.000000
OC         count     16.000000
           mean      35.625000
           std       64.555790
           min        0.000000
           25%        1.000000
           50%        8.500000
           75%       23.250000
           max      212.000000
SA         count     12.000000
           mean      62.416667
           std       88.620189
           min        1.000000
           25%        3.000000
           50%       12.000000
           75%       98.500000
           max      221.000000
dtype: float64

Step 6. Print the mean alcohol consumption per continent for every column¶

In [10]:

drinks.groupby('continent').mean()

Out[10]:

	beer_servings	spirit_servings	wine_servings	total_litres_of_pure_alcohol
continent
AF	61.471698	16.339623	16.264151	3.007547
AS	37.045455	60.840909	9.068182	2.170455
EU	193.777778	132.555556	142.222222	8.617778
OC	89.687500	58.437500	35.625000	3.381250
SA	175.083333	114.750000	62.416667	6.308333

Step 7. Print the median alcohol consumption per continent for every column¶

In [14]:

drinks.groupby('continent').median()

Out[14]:

	beer_servings	spirit_servings	wine_servings	total_litres_of_pure_alcohol
continent
AF	32.0	3.0	2.0	2.30
AS	17.5	16.0	1.0	1.20
EU	219.0	122.0	128.0	10.00
OC	52.5	37.0	8.5	1.75
SA	162.5	108.5	12.0	6.85

Step 8. Print the mean, min and max values for spirit consumption.¶

This time output a DataFrame¶

In [15]:

drinks.groupby('continent').spirit_servings.agg(['mean', 'min', 'max'])

Out[15]:

	mean	min	max
continent
AF	16.339623	0	152
AS	60.840909	0	326
EU	132.555556	0	373
OC	58.437500	0	254
SA	114.750000	25	302