What does a PhD student do when given a gift card to go shopping? Ask for clothing recommendations and analyze the results!
The setting: I had a gift card of about €200 to the Swedish department store Åhlens. Given my non-existent sense of fashion, I asked my fiancée to give me clothing recommendations. She sent me an Excel file with each row containing a link to an item and a personal note on why she likes this piece. Each sheet was for a different category of items (e.g. long sleeves, sweaters, etc.). You see why I love her, right?
So, I wrote a short Python script that for each item, scrapes the following info: brand name, item name, color, online storage status, warehouse storage status, price, rating, and description. This information wasn’t available for each item. With the most available information, here’s what I discovered about her recommendations for me.
Most clothing items were for around 1,000 SEK (~€99). The price distribution is as below.
Most of her favorite items were long sleeves, followed by short sleeves, shirts and jackets.
If I were to go around the department store myself, where would I start? I grouped the items by their brand name. The clear winner is ‘John Henric’.
Now I know which brands she likes on me. The next question is, what colors should I go for?
I’m not a fan of complicated color names (what is moonstruck or mole mel??). I took the fancy names and converted them to ‘regular’ colors. I took each color and plotted it in proportion to its frequency in the data. The results were clear: I should wear greens and browns.
The perfect clothing item would be a long sleeve from John Henric in either green or brown, and might cost around €90.
If I don’t buy an expensive item, I could also optimize the use of the gift card by buying the greatest number of items while sticking to my budget. So I wrote a simple function that finds how many (and which) items I could buy within a given budget. For example, for 3,000 SEK or approx €300, I could have bought all of the following:
Overall, shopping has become much more enjoyable now that I can run some code on it.
Uploaded 2025–06–10; Last updated 2025–06–10