My team proposed a study that would analyze runway fashion to track overall trends in the fashion industry because these trends are generally started by high fashion houses on runways. We would gather data by scraping pictures of outfits in runway shows from the archive on style.com, which stores pictures of each outfit in a fashion show separated by designer and season.
In order to track trends, we would do image processing on all the collected pictures and attempt to gather meaningful information. Because each picture will have a similar background and the models will be standing in similar positions, the data set would allow for easy detection of trends between designers because we could check for what sort of colors or shapes (big sunglasses or smaller ones, etc.) occur frequently in clothes while ignoring the background.
This would be helpful for consumers because they would be able to see what will become popular in the fashion industry before it actually happens.