W207_Group_Project

W207 Summer 2018 Group Project

The data have been pre-processed:

• keep only exercises that can be classified into big 3 lifts (squat, bench and deadlift)
• remove invalid sets and reps
• unroll nested json

Sets data:

• filename: set_data_w207.csv
• unique on setID
• contains data relevant to each set (weight, exercise, RPE etc.)

Reps data:

• filenames (based on different app releases):
  • rep_data_w207_new.csv (contains 22 columns)
  • rep_data_w207_old17.csv (contains 17 columns)
  • rep_data_w207_old19.csv (contains 19 columns)
• unique on setID + RepCount (RepCount starts with 0)
• contains data relevant to each rep (average velocity, range of motion etc.)

How to combine datasets:

1. Stack reps data (order is not important)
2. Merge with sets data by setID

Caveats (ideas for future processing):

• Some numeric fields are strings (weight, RPE)
• Those strings can have , or . to indicate fractional numbers
• Weight can be in kgs or lbs (use metric column to determine which one it is)
• Weight and RPE can have extreme (invalid) values that need to be removed
• Need to create cleaner labels for exercises (and potentially remove sets with tags)