I'm currently coding a program to automatically place bids and buy reasonably priced unpopular products. You see, I finished university with a degree in computer science with specialization in machine learning. I thought this would be a good idea to use a deep neural network to automate purchasing. The reasoning being is that I can make some nice profit but at the same time increase the liquidity on the site making it better for sellers. In other words, sellers would sell more products and sell them quicker.
I'm currently thinking of making a multimodal neural network that takes in tabular data related to the listings( e.g. price, condition, title...etc.) and corresponding image data to then come up with a price prediction. In order to do this I need access to historical data. I.e. how much similar items have sold for in the past. I understand you can do this with the marketplace insights api but there is no where to to gain access to it. And so I resorted to writing here hoping that you see some good potential in this project. I'm happy to provide more information if you need it.