Column

Methodology

• The dataset for this analysis is obtained from Tidy Tuesday’s 2021-12-21 edition on Starbucks drinks.

• This report will only use the data collected by Tidy Tuesday it has no contribution or input to the methods used to generate the data. It will only utilize the data as given, anything outside of the dataset is beyond the scope of this report.

Limitations

• As the data is exogenous to this report, we cannot be certain that the information on Starbucks drinks are completely representative of their real life equivalent.

• Some product items data are complete duplicates of each other in terms of add-ons, and size but the nutritional information differ, this may add additional uncertainty to the report.

Processing the data

• The unaltered data is ‘starbucks’

• Since there are duplicates in the ‘starbucks’ data we will first seek to eliminate the duplicates. It is considered a duplicate if the name, milk, whip and size is the exact same, and we need to remove it from the data. (‘starbucks_M’)

• Some of the variables (fiber and trans fat) are categorized as characters when it should be numeric. Thus, we will convert them into the appropriate data type. (‘starbucks_M’)

• On the menu there are extras which have a serving size of zero, such as additional shots of espresso. Since we are only interested in drink items; we will remove these observations. (‘starbucks_M2’)

• In order to build our linear model, we will need to standardize the size of each drinks so that the nutritional content is reflected in a comparable amount of volume. (‘starbucks_standard’)