Tracking 2018 (March) - Outfits

Back to project

For March, I tracked the outfits I wore. I took an "Outfit of the Day" photo every day, and then quantified and analyzed them according to article of clothing, brands worn, and tried to extrapolate trends from my daily aesthetic. Please note that the variety is restricted due to my own wardrobe choices and being away at school.

The graphic above depicts a general overview of the colour combinations I wore during the month. The first row is for days 1-8, the second row for days 9-16, then days 17-24, and lastly, days 25-31.

Not surprisingly, I wear a lot of black. I expected to wear something black every single day (my go-to is a pair of black jeans); however, I didn't wear black at all on March 24. I wear a limited selection of colour, but I do cycle through blues, reds, and pinks quite regularly (a bit more than expected).



I clustered my dataset using R, and plotted them on a graph. The components/dimensions on which my cluster plot was graphed on were based on system-determined similarities. From this plot, 5 trends were observed: blue jeans, a cardigan + tee combo, monochromatic + belted combo, plaid button-down days, and other combinations.

Because March has 31 days, I split my "weeks" into segments of 7-8 days for a more even analysis.

My most popular brands were by far Muji and Uniqlo; I would wear them 4-6 days out of every 7-8 days.

I'm a fan Adidas, but to my surprise, I didn't wear it that often.

Other brands include Zara, GAP, Jcrew, etc. They were grouped in "other" because alone, my frequency of wearing them was insignificant.

I calculated a few regression models to observe the predicability of certain outfits or trends:

  • The likelihood of wearing black bottoms if I was wearing a black top was 90.32%.
  • There was a 45.26% chance I was wearing Adidas, Zara, Jcrew, or other brands if I wore a sweater that day.
  • The likelihood of wearing Muji and Uniqlo together was 74.2%. Also, my pants were 90% likely to be from Muji or Uniqlo.


I plotted a scatter plot to visualize the types of belts I wore with the types of outfits (I used all-black as a baseline to compare to). As it turned out, black stuck to black (or no belt), and everything else was relatively fair game.

Previous

February

Next

April

Project Home

Tracking 2018