Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

geospatial use case #121

Open
flutter-painter opened this issue Oct 23, 2024 · 2 comments
Open

geospatial use case #121

flutter-painter opened this issue Oct 23, 2024 · 2 comments

Comments

@flutter-painter
Copy link

Hi @pamelafox,
Congratulations for providing such a clear and comprehensive demo with a common use case.

Yet in this common retail scenario, I see little added value over a classic ecommerce search view.
The keywords 'climbing' and 'gear' would match just the same with full text search over title/description.
The price range could be typed by user using one text field max price : 30
While the conversational UI gives this "communication feeling", it is not handy to display dozens of matching products.

I wonder if other use cases would illustrate the unique querying capabilities of OpenAI.
Do you think it could handle geocoordinates, or even postgis ?
This would really change the experience

@pamelafox
Copy link
Contributor

I think that'd be quite doable!
If I was going to add that, I would start with a table with lat/lng coordinates. Then when the user asked the question, I'd use the query rewriting step to pull out an address/location string. Then I'd send it to the Bing geocoding API, and use it with the PostGIS operators along with the vector/text search.

I don't know when I'd have time to try it out, but it sounds pretty compelling and doable.

Is that what you were thinking?

@flutter-painter
Copy link
Author

flutter-painter commented Oct 24, 2024

This is encouraging !
Well I had nothing particular in mind, it just seems there are lots of efficients ways to search for products, but few when it comes to combining products & locations.

You could stick to this retail scenario but twist it for a company use-case, for example ; help a salesman find available stock for a product P across dozens of stores.

You could spice it up a bit by linking each stores to multiple stockrooms and split the stockrooms in different categories (remote, damaged, staff sales only). There it would make a lot of sense to ditch the GUI for a chat.

I don't know what is possible with this this sample PostGIS dataset (countries and capital), but I presume if you pin the stores in a few well-known capital the requests could become really valuable, e.g. is there any product P available for staff-sales in New-York or Ottawa ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants