search

mixnjuice / project-ideas

This is an attempt to create the best, most user-friendly mixing website the world has ever seen :)
9 stars 2 forks source link

Data mining #14

Open pscn opened 5 years ago

pscn commented 5 years ago

Another person really wanted to see some historical data which should be an amazing tool to have. Inventory prices, juice prices, which flavour do I use often, etc.

That's something I also found intriguing. How much of the flavors am I actually using? Why is this flavor already empty? How much did it cost the last time I bought it and from where did I get it again?

A solution to this might be to write (a) log table(s) that we can later query for all kinds of cool reports. I'm not sure if we want to track the shops where we bought stuff though. That adds another layer of complexity.

As a side note: another thing I would love to know is how much liquid I vape daily / weekly / monthly. To track that I was thinking about weighting my bottle after mixing and than weighting it again every few days to calculate how much I actually used. If we'd have a feature like that and can get that data it would be mind blowing. Top 10 of weekly, monthly recipes not only by how much it was mixed, but also by how much it was vaped. However: at the end of the day I don't think too many people would use it and the little data we'd have would be useless. But just wanted to through out the idea to get some feedback on what you guys think about it.

Korlimann commented 5 years ago

I think this is an amazing idea. Although you are right, it might be that people are not really interested in that. If they are though, we could get so much interesting data. One thing that has been interesting me for some time is how much ml I vape on a daily basis. Something where I could calculate that, and keep track of my usage and costs is what I've been wanting for some time. We could not only get data for the website, but also general data, as in "How much ml does the vaping community vape on a daily basis on average?" and the like. If we have time for that, that's something I'd really like to implement as well.

Humanpuck commented 5 years ago

This is a great idea and you're only scratching the surface of what you could do with the data. Once the data starts to build you can offer things like "most used Apple flavors," which could double as a new user tool. NLP to analyze descriptions and compile "flavor facts" things like that.

In terms of storing data I would assume you're going relational for things like flavors/vendors/suppliers etc. and storing description as free text maybe BLOB?

pscn commented 5 years ago

AFAIK it's not decided yet what we use for the persistence layer. I think we should collect the requirements first and than decide what the best tools for the job are. Right now I only created the "feature" issues based on what @Korlimann collected in the wiki. We should also start issues discussing specifically what we want to store and what kind of data we expect.