• The following is the Python libraries we need for this homework.
import pandas as pd
import seaborn as sns

Question 1

Load the CSV file, beer_markets_cleaned.csv, from the class website:

beer_markets = pd.read_csv('https://bcdanl.github.io/data/beer_markets_cleaned.csv')
  • Description of DataFrame beer_markets
    • One single observation (row) corresponds to information regarding one household and its beer purchase.
    • DataFrame beer_markets does not have any missing values.


  • Variable (column) description
    • hh: an identifier of the household;
    • purchase_desc: details on the purchased item;
    • quantity: the number of items purchased;
    • brand: Bud Light, Busch Light, Coors Light, Miller Lite, or Natural Light;
    • spent: total dollar value of purchase;
    • beer_floz: total volume of beer, in fluid ounces;
    • price_per_floz: price per fl.oz. (i.e., beer spent/beer floz);
    • container: the type of container;
    • promo: Whether the item was promoted (coupon or otherwise);
    • market: Scan-track market (or state if rural);
    • demographic data, including gender, marital status, household income, class of work, race, education, age, the size of household, and whether or not the household has a microwave or a dishwasher.



Q1a

  • Sort the DataFrame beer_markets by hh in ascending order.


Answer

beer_markets = beer_markets.sort_values('hh')



Q1b

Count the number of households for each market.

Answer

# The below line counts the number of non-missing values
# for each variable for each group of "market" and "hh".
q1b = beer_markets.groupby(["market", "hh"]).count() 

# In the resulting DataFrame q1b above, 
# there is only one single observation for each group of
# "market" and "hh".
# So, the following line counts the number of households
# for each market.
q1b = q1b.groupby(["market"]).count()

q1b = q1b.loc[ :, ["brand"] ]
q1b.columns = ['n']   # rename `brand` with `n`

# For renaming, the following also works:
# q1b = q1b.rename({'brand':"n"}, axis = 'columns')



Q1c

Find the top 5 beer markets in terms of the number of households that purchased beer.


Answer

q1c = q1b.sort_values('n', ascending = False)
market n
TAMPA 280
DETROIT 214
COLUMBUS 184
MIAMI 178
PHOENIX 168



Q1d

Sum of beer_floz for each market.

Answer

q1d = beer_markets.groupby(["market"]).sum()
q1d = q1d.loc[ :, ["beer_floz"] ]

# the following gives the Series, instead of the DataFrame
# q1d = q1d['beer_floz']



Q1e

Find the top 5 beer markets in terms of the amount of total beer consumption.

Answer

q1e = q1d.sort_values('beer_floz', ascending = False)
market beer_floz
TAMPA 462904
PHOENIX 388824
DETROIT 309588
MIAMI 271870
COLUMBUS 266896



Q1f

  • Variable price_per_floz is continuous.
  • Variable brand is categorical.

Describe the distribution of price_per_floz for each brand using seaborn.

Make a simple comment on comparison for the distribution of price_per_floz across brands.

Answer

# distribution of price_per_floz for BUD LIGHT
sns.displot( x = 'price_per_floz', bins = 200,
             data = beer_markets[ beer_markets['brand'] == 'BUD LIGHT' ] )

# distribution of price_per_floz for COORS LIGHT
sns.displot( x = 'price_per_floz', bins = 200,
             data = beer_markets[ beer_markets['brand'] == 'COORS LIGHT' ] )

# distribution of price_per_floz for MILLER LITE
sns.displot( x = 'price_per_floz', bins = 200,
             data = beer_markets[ beer_markets['brand'] == 'MILLER LITE' ] )

# distribution of price_per_floz for BUSCH LIGHT
sns.displot( x = 'price_per_floz', bins = 200,
             data = beer_markets[ beer_markets['brand'] == 'BUSCH LIGHT' ] )

# distribution of price_per_floz for NATURAL LIGHT
sns.displot( x = 'price_per_floz', bins = 200,
             data = beer_markets[ beer_markets['brand'] == 'NATURAL LIGHT' ] )


  • BUD LIGHT, COORS LIGHT, and MILLER LITE have the similar distribution of price_per_floz with each other.

  • BUSCH LIGHT and NATURAL LIGHT have the similar distribution of price_per_floz with each other.

  • Overall, BUD LIGHT, COORS LIGHT, and MILLER LITE are more expensive than BUSCH LIGHT and NATURAL LIGHT.



Q1g

  • Both variables price_per_floz and beer_floz are continuous.

  • Describe the relationship between price_per_floz and beer_floz by brand using seaborn.

  • Make a simple comment on the visualization result regarding how the relationship between price_per_floz and beer_floz varies by brand.


sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           scatter_kws = {'alpha' : 0.2},
           data = beer_markets[ beer_markets['brand'] == 'BUD LIGHT' ] )


sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           scatter_kws = {'alpha' : 0.2},
           data = beer_markets[ beer_markets['brand'] == 'COORS LIGHT' ] )


sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           scatter_kws = {'alpha' : 0.2},
           data = beer_markets[ beer_markets['brand'] == 'MILLER LITE' ] )


sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           scatter_kws = {'alpha' : 0.2},
           data = beer_markets[ beer_markets['brand'] == 'BUSCH LIGHT' ] )


sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           scatter_kws = {'alpha' : 0.2},
           data = beer_markets[ beer_markets['brand'] == 'NATURAL LIGHT' ] )


# all the plots in one plot:
sns.lmplot(x = "beer_floz",
           y = "price_per_floz",
           hue = "brand",
           scatter_kws = {'alpha' : 0.2},   
           data = beer_markets)


  • The law of demand holds in the beer market:

    • From the fitted straight line, we see the downward-sloping demand curve for each beer brand.

  • The demand curves for BUD LIGHT, COORS LIGHT, and MILLER LITE are steeper than those for BUSCH LIGHT and NATURAL LIGHT.

  • According to the demand curves, BUD LIGHT, COORS LIGHT, and MILLER LITE have larger demands than BUSCH LIGHT and NATURAL LIGHT given the same level of price_per_floz \(\geq\) 0.0375.