Inspiration

Our teams shared passion for learning and geography.

What it does

Provides visual context to written content. Enhances comprehension, engagement, and retention. Users can search for any query

How we built it

We drafted the prompt for the language model while researching how to build a base map. Given the amount of time we had, we used living atlas to achieve a simple design. We then created a Flask Rest API which is the server that interfaces with Open AI's api to construct a response for the users query and then sends backs content (LLM response, events with geographic data + images) to display on the map. Our web app uses ArcGIS maps SDK for JavaScript to dynamically display the content on the web map. We also added 2 basemaps (satellite & modern boarder)and front end custom components, feature layers, and points that are coming from the back end server.

Challenges we ran into

Some challenges we faced include the process of iterating through different versions of what we wanted, adjusting the project based on the available data and functionality we can actually create within the time frame, prompt engineering and refinement, geocoding inaccuracies, the inability to view the full map to show all of the points, getting the base map to connect to the web maps, and the Adobe After Effects learning curve.

Accomplishments that we're proud of

Aylee: The dedication and effort put in by everyone, as well as our ability to use our diverse skill sets to create something we can be proud of. Wesley: I was able to bridge the gap between the technical and design aspects of our project and have a versatile role in the group. Alicia: Creating our team's project presentation and Devpost submission, while also playing a versatile role within the group. Abhi & Gaurav: We are able to utilize Esri technology to accurately display points, use geocoding, scrape images, and incorporate most of the features we initially planned. Our user interface is user-friendly, and we're especially proud of our image feature.

What we learned

Power of friendship!!! Creating a interactive web map in 2 days is hard but we learned a lot about Esri technology (e.g. ArcGIS maps SDK for JavaScript, Esri's geocoding service through ArcGIS API for Python, Web Maps, etc.).

What's next for Latitude with Attitude

  • Using Esri's python libraries for NER from unstructured text
  • Feature to export knowledge graph based on user's exploration
  • Adding Guardrails to the LLM responses used to generate paragraph responses to prompts
  • Perhaps connecting to database of historical articles / a museum archive
  • Nice to have features (e.g. save/export as Feature Service)

Built With

Share this project:

Updates