In 2017, I launched Three Six Five Closet and spent the last few months of the year creating database with details about every piece in my closet. From 2018 to 2020, I (attempted) to document every single item I wore every day, and assigned each outfit a sentiment score. What began as a project to determine the cost per wear of every item I purchased (to see if I was truly buying “investment pieces”) turned into a search for my personal style and a love of fashion.
I no longer maintain a database of every outfit I wear. While I stuck to the exercise for almost two years, I would feel guilty when I inevitably forgot to track a day or two. When Covid-19 hit, I stopped the practice altogether. I realized I no longer needed to look to data to help me understand if an item suited my personal style.
Today, I still maintain an inventory of my entire wardrobe. I find this exercise to be much more manageable and informative. Since 2017, I have tracked fields such as date of purchase, cost of item, color, size, reason for purchase, etc. Instead of analyzing every outfit on a daily basis, I analyze my closet data on an annual basis to look for patterns in why I purchase items and why some items I eventually let go of.
My closet and personal style have continuously evolved since 2017. I used to think there would come a point at which my closet database would be “production ready”, whether that meant owning only a very low number of items or not buying anything for a few years.
Now, I know the day may never come when my wardrobe is a perfectly curated capsule. I own pieces I wear regularly that don’t “spark joy”, and I’ve held onto pieces I never wear that do. I’ve experienced fluctuations in my body that have made never shopping an unrealistic standard.
I started this project in 2017 ultimately to better understand myself through my closet. Nowadays, I use this dataset to evaluate my reason for purchasing new items. Fittingly, as the pieces in my closet have changed over the years, so have the answers I try to seek from my closet data.

