Project 2 Approach - Tidying Three Wide Datasets

Introduction (Approach)

For this project, I will tidy and analyze three wide datasets from our class discussion: U.S. inflation rates (Samantha Mesa), population estimates for the 10 most populous states (Anastasiia Gmyrina), and average U.S. tuition from 2004 to 2016 (Mubin Ejaz). Each one is wide for a different reason, so I expect the tidying steps to vary from dataset to dataset.

U.S. inflation rates. The source table on the U.S. Inflation Calculator website has one row per year, a separate column for each month, and an annual average column next to the monthly values. I plan to recreate this table as a CSV, remove the annual average column since it can be recalculated from the monthly values, and reshape the months into a single column so each row represents one year and month. The main challenge is that some cells contain text, such as “Avail. Sept. 11,” instead of a number. I will need to convert those to missing values before treating the column as numeric. Once the data is tidy, I would like to compare inflation before, during, and after the 2021 to 2022 spike, and check whether the annual average hides large month to month swings.

10 most populous states. This is the simplest of the three, with one row per state and a column for each year from 2020 to 2022. I plan to reshape the year columns into rows, then calculate the percent change in population from 2020 to 2022 to see which states grew and which ones shrank.

U.S. average tuition. This dataset comes from TidyTuesday and has one row per state, with each academic year in its own column. The year labels are stored as ranges like “2004-05,” so I will need to decide how to convert them into a usable year value. The tuition amounts may also be stored as text with dollar signs and commas, which would need cleaning before any calculations. Once tidy, I plan to look at how tuition grew over the 12 years and which states saw the largest increases.

For all three, I will keep each raw file in its original wide format in my repository before tidying, so the starting point is clear and the cleanup steps can be reproduced.