search

cmoog / traderjoes

Price tracking program and website for Trader Joe's, updated daily.
https://traderjoesprices.com
MIT License
201 stars 15 forks source link

Expansion of current site and maybe more? #8

Open undead9786 opened 1 month ago

undead9786 commented 1 month ago

This is a great resource and I tried myself to scrape the site but no such luck on my own so curious to see if there can be improvements done.

1) Having a search button would be a great feature. I know you can just do a find but with the amount of stuff on here it would be cool if you can do a search to see filtered choices to see the prices better.

2) Adding in the weight and also a price per weight comparison. With everything constantly going up people would find it beneficial to have the ability to see their best choice of value.

3) Personalize seeing stores of your choosing. Being able to add the store close to you and prioritizing those prices over everything so you can see if locally the prices are different. New to TJ's so not sure if prices are the same across all stores are if there are fluctuations.

4) Adding manual entries. TJ's doesn't post everything they provide in store like whole milk so would be great if you can manually add an entry.

5) If possible to make it so we can run this locally ourselves would be awesome as an option.

6) Adding a rating system. Would be great to have a rating system so people can show what are good things to try and potentially what to stay away from.

7) Has there been any thoughts about expanding this? More and more stores are online and would be great to potentially collect from as many sources as possible and potentially have the best method of choices on what to purchase and from where.

cmoog commented 1 month ago

I don't have any immediate plans to increase the scope of the site. This is largely why I've chosen the MIT license and keep the raw dataset available at https://data.traderjoesprices.com/dump.csv. I'll share my thoughts on your suggestions below, but as always, feel free to fork if you're motivated to implement them yourself.

  1. This would be a nice UX improvement.
  2. A benefit of the current approach is that the data is extremely sanitary. sku and price come directly from the product catalog, have identical semantics across all products, and are unlikely to change in meaning over time. Not all products are priced or bundled by weight.
  3. The raw dataset includes 4 store locations. I may at some point add a selector for these stores in the UI. https://github.com/cmoog/traderjoes/blob/156f52f470ca76b01c0e18e07ee4c0d97f15d4be/Main.hs#L51-L54
  4. Similar to point 2, I have a strong preference for data cleanliness. Manual entries preclude this.
  5. This is already possible today. I include build steps in the readme using the nix build tool, which encodes all necessary dependencies for building and running the program from source.
  6. I'm not interested in growing the scope to include social or crowdsourced features.
  7. I'm interested in "expansion" only insofar as that entails aggregating high quality datasets through automated collection of retail catalogs, not implementing social features or otherwise decreasing dataset quality with manual additions.
undead9786 commented 1 month ago

Yea still all new to github workings and any type of coding outside of batch scripts and powershell.

  1. =)
  2. Gotcha, also noticed that lack of weight thing too and they go 1 item or something dumb.
  3. Yea saw that in the main.hs, any reason why you are scoping to those 4 spots? Are those like the main hubs that you can potentially get the most amount of products that may not be available on the site of a specific store but can be in the store?
  4. Understood.
  5. Not familiar with the nix build tool, is it just a program similar to docker that just runs packages? Will need to look more into that.
  6. Got it, maybe if I run your package locally and learn how it works I can make changes to it to support some of the stuff I would personally be looking for =)
  7. For this one I wasn't actually looking for social features or anything. More like taking the data off of multiple different stores and aggregating them into 1 so you can see the differences in prices of items rather than going to each site or delivery service and hunting down items for each thing. Such as having TJ's, Whole Foods, Instacart (of a specific store), Morton Williams and being able to search for something like watermelon to see prices in all locations in 1 spot.