Applying Data Analysis to Personal Wardrobe Optimization

Every morning the author faces the dilemma of choosing what to wear, which often leads to indecision and wasted time. Inspired by experience in data analysis and recommendation systems, she proposes to solve this personal problem by applying a systematic data‑driven approach.

The problem is defined as improving the "clothing pool" (the wardrobe) and the decision‑making strategy for purchasing new items. The goal is to identify why the existing wardrobe feels insufficient and how to optimize it.

Data collection involved photographing every piece of clothing and labeling each item with attributes such as type (t‑shirt, sweater, dress, etc.), season, purchase time, purchase channel, color, distinctiveness, and wear frequency. Dirty data (gifts, special‑occasion items that are never worn) were removed to ensure clean analysis.

Key metrics were defined:

Quantity – total number of items and their distribution across seasons, types, and channels.

Elimination rate – the proportion of items that are "never‑wanted to wear" over the total wardrobe.

Analysis of quantity showed that summer clothing is most abundant while winter items are scarce, matching climate expectations. Frequency distribution is left‑skewed, indicating many rarely worn pieces. Cross‑tabulations revealed imbalances such as 11.5 tops per bottom in spring and a shortage of versatile denim pants.

The elimination rate averages around 30 %. Seasonal breakdown shows higher elimination in winter, while summer items have lower rates, partly because recent purchases dominate summer wear. Purchase channel analysis indicates that online‑bought clothes have a higher elimination rate than gifts or store‑bought items. Style analysis suggests that highly distinctive pieces are more likely to be discarded, whereas conservative items are safer choices.

Typical bad cases were identified and annotated, providing concrete examples of why certain items are abandoned and suggesting corrective actions.

Based on the findings, the author proposes several actionable strategies:

Prioritize buying denim pants.

Try on winter clothing in physical stores to avoid unsuitable purchases.

Maintain the current summer wardrobe; occasional online additions are acceptable.

Avoid flashy spring pieces that tend to be unused.

Promptly return ill‑fitting online purchases, as online shopping is the top cause of elimination.

Finally, the author emphasizes the importance of continuously updating the dataset, monitoring the defined metrics, and iterating the strategy to keep the wardrobe ecosystem healthy.