search

codeforUV / food_resources

A project that aims to increase food access for residents of the Upper Valley region of VT/NH
https://food-resources.herokuapp.com
MIT License
1 stars 6 forks source link

Consider Open Referral's Data Specification in our data model #64

Closed thomascchen closed 2 years ago

thomascchen commented 3 years ago

Open Referral's website: https://www.openreferral.org

Open Referral’s goal is to enable interoperability among diverse technologies that might use resource directory information to help meet people’s needs.

They are working on a Human Services Data Specification and a set of protocols for "accessing, updating, exchanging and synchronizing community resource directory data."

Perhaps their standards can inform how we model our data?

thomascchen commented 3 years ago

I reviewed the HSDS in more detail today, and I'm really impressed by the level of thought and consideration that's evident in this work. I think the specification lends itself very well to the structuring of data in our food resources app. Reading through the specification made me aware of a number of things that we had not really considered or had glossed over in our initial thinking, particularly accessibility and language data. The design of the schedule table is also extremely helpful as figuring out how to represent schedules is something that has been vexing me. I think what I will do next is to design/implement a local implementation of this with the small data set that we have. I will likely omit a lot of tables that are not relevant but try to stick to being HSDS compliant. The only thing I think we might need to "extend" is the concept of "recently available items".

If there is ever a centrally hosted HSDA, we could proxy most of our requests to that while maintaining a local database for "recently arrived items". In the shorter term, it may make sense for us to start with our own local database modeled in an HSDS compliant way, then evolve later to perhaps synchronizing with a centrally hosted HSDA via background or scheduled jobs, and then eventually maybe we could migrate to straight up proxying all requests straight to the HSDA. Those are just some initial thoughts.

iritush commented 3 years ago

Thanks for the great summary @thomascchen. This all makes sense. Let us know how it goes and if you need any feedback

thomascchen commented 2 years ago

@iritush Took me a long time, but I did give the HSDS data model a good try in this PR: https://github.com/codeforUV/food_resources/pull/78

Ultimately, I found it a bit too cumbersome to work with for our current needs, so ended up reverting it, but happy to provide more feedback if interested.

iritush commented 2 years ago

Thank you, @thomascchen. Very much appreciated! We are not actively working on this project at the moment but will reach out to you when we work on it again