search

QMSS-G5063-2022 / Group_H

0 stars 0 forks source link

Submitting project proposal #1

Open franyang6 opened 2 years ago

franyang6 commented 2 years ago

@QMSS-G5063-2022/teaching_team SHA: 57350c4b6fa26b5936416f90d7fa2cac4f73a250

Rishav1999 commented 2 years ago

Based on your proposal, you will be understanding the clothes trend based on apparels sold by e-commerce sites. The datasets you have taken are the set of apparels sold by these companies. The question to ask is: 1) how can you claim that if an e-commerce is selling a type of clothing, that is a trend during that time? Is there any data to support that? 2) how are you sure that you can realize the sentiment for a particular category of clothing just by scrapping Twitter Data and not say Instagram where you might find more fashion trends? Is there any data to support that?

JonathanReeve commented 2 years ago

Just to add my two cents to this one: be careful when quantifying colors, especially if they're color words, since these are often difficult to categorize. If it were me, I'd want to automatically extract the color of a product from its image, using, e.g., color-thief.py. That way, you're not beholden to the categories or descriptors in the dataset.