Inspiration
Yesterday, my friend sent me a message in Facebook because she remembered me out of nowhere. She is an expert in GIS. We talked a lot about GIS, tools and AI. Then, this morning, I saw a Facebook post about the WebMCP challenge. Thus, I have thought of this idea.
What it does
Basically, this is a web app featuring a geo dataset. For this demo, I chose just 1 sample public dataset from Melbourne datasets. This dataset is just sample data from the playgrounds dataset. In usual web apps, data or map is displayed. Then, users will query the data based on the filters they need. Example: Which playgrounds have swings or sandpit? Think of niched micro sites or niched directories.
As required in the WebMCP challenge, WebMCP is added in the web app. With this, the user puts the question to ChatGPT in natural language, then ChatGPT calls the tools exposed in the web app using the WebMCP. Thus, the user does not need to do the filtering in the web app. ChatGPT does the searching for the user.
Imagine the possibilities if we add this kind of functionality in web apps or dashboards with a lot of GIS datasets. Business users and managers do not need to struggle using complicated dashboards. They just need to "talk to their data" using ChatGPT as the digital worker.
How we built it
I used Codex to write the code. I tested the app myself.
Challenges we ran into
I had little time for building. Because of this, I was not able to think well what is the better use case.
Accomplishments that we're proud of
I was able to understand how WebMCP works. I got more interesting ideas that I can implement after this simple demo project.
What we learned
WebMCP works, Codex works. Melbourne and Victoria government have interesting datasets that can be put into an app
What's next for MapsBrief
Rethink of what's better product features such as adding different datasets that are useful for more business insights.
Built With
- cloudflare
- codex
Log in or sign up for Devpost to join the conversation.