As a test I searched for "switch", I was being intentionally vague, but I meant "Nintendo's Switch Console". The switch query brings up a bunch of switch related accessories and games, which was ballpark what I was looking for, but not quite.
I tried "Nintendo Switch" next, which brought up a bunch of different versions of the Switch, but Walmart kinda overwhelmed other marketplaces. Although two did pop up from Best Buy in the second page for me.
I tried "Switch Console" next, and it was 100% Walmart listings for at least three pages.
The sites still useful as it, it just takes a little digging to get what ya want. Search is hard as hell, it's understandable. As this evolves, it would be nice if you tried to normalize products across different stores and showed them in aggregate, so if I search "Nintendo Switch" I'll get results from every store that carries the "Nintendo Switch OLED, White Model" together. It'd be super easy to track the cheapest price for the specific thing I'm looking for over time that way.
Did you see the post about the release of CozoDB 0.6 a little while back? I'm not affiliated with the project, but building product ontologies was one of the examples he gave. Maybe check it out? https://docs.cozodb.org/en/latest/releases/v0.6.html
I am definitely hoping / planning to offer that unified "view this product across stores" feature — that would be killer. Just have to work on the ML stuff to get that sufficiently good and get the data in a good spot. The CozoDB thing looks very promising, will check it out! Thanks again
I'm a staff software engineer by trade, but love working on side projects in my (relatively rare) spare moments. This has been a fun project, involving some new things I hadn't worked on before (extended parsing / crawling, scraping, headless browsers etc.). Hoping to put together a write-up on the tech soon.
