Exploratory Data Analysis on Beers & Breweries Datasets

Personal Projects #Data Science#Python

Data Cleaning#

Merge the two datasets on brewery_id and id to create a joint dataset. Then, determine the column categories needed in a dataframe by identifying unique columns. The categories are:

Column	Type
abv	numeric
ibu	numeric
id	numeric
name	categorical
style	categorical
brewery_id	numeric
ounces	numeric

Descriptive Statistics of Numerical Values#

Numerical Values in Columns#

Determine the number of numerical values in each column for the series by finding the length of each column. brewery_id and id were not included.

Column	Length
ibu	2410
abv	2410
ounces	2410

Non-Null Values#

Find the number of non-null values in each numerical column (brewery_id and id not included) by using count().

Column	Non-Null Values
ibu	1405
abv	2348
ounces	2410

Then, determine the percentage of null values in each numerical column using the following calculation:

((length_of_column - count_of_missing_values)/length_of_column)*100

Column	Percentage of Missing Values
ibu	41.7%
abv	2.6%
ounces	0%

Minumum & Maximum Values#

Min and Max values of each numerical column (brewery_id and id not included) are determined using min() and max().

Column	Min	Max
ibu	4.0	20.0
abv	0.001	0.128
ounces	8.4	32.0

Mean, Median, Mode, & Standard Deviations of Numerical Columns#

Mean, median, mode and standard deviations of each numerical column (brewery_id and id not included) are determined using the following functions: mean(), median(), mode(), std().

Column	Mean	Median	Mode	Standard Deviation
ibu	42.713	35.0	20.0	25.954
abv	0.0598	0.056	0.05	0.0135
ounces	13.592	12.0	12.0	2.352

Quantile Statistics#

Quantile statistics are used to determine data spread, skewedness and outliers. To determine quantile statistics, split the data into equal sized groups using cut points (quantiles). For example: beers['ibu'].quantile([0.25,0.5,0.75])