Foursquare NYC Venue Search

NYC semantic venue search (Streamlit + Snowflake)

Description

Discover NYC with this Streamlit app! It utilizes cosine similarity to semantically match your search query with Foursquare venue categories, helping you find matching venues in your selected neighborhoods across the city. Powered by Streamlit, Snowflake, and OpenAI, this app allows users to explore NYC like never before.

Features

Select up to five neighborhoods in a borough to focus your search
Input a search term (e.g., "Epic night out") to find relevant venue categories
The app uses cosine similarity to match your query with Foursquare venue categories
View suggested venues on a map and in a table with venue details

Technical Implementation

Streamlit app for the frontend user interface
OpenAI API to generate text embeddings for search queries and for Foursquare venue categories.
Snowflake for data storage and retrieval

Building the App

Follow these steps to build the app:

Clone the repository.
Follow the steps in /scripts/snowflake_setup.sql to embed the Foursquare venue categories and set up the database and tables in Snowflake.
Create a .streamlit/secrets.toml file containing the following:

[snowflake]
user = ""
password = ""
account = ""
warehouse = ""
database = ""
schema = ""

[openai]
api_key = ""

Accessing the App

You can access the app on the Streamlit Cloud community at nyc-venue-search.streamlit.app. Please note that the hosted version of this app may be taken down at any time, as it was created for the Streamlit Summit Hackathon. To build and run the app yourself, you can clone the repository, set up Snowflake, and follow the steps in the "Building the App" section.

Feedback

If you have any feedback or questions about this app, please contact me on Twitter at @dclin. Thank you for checking out the tool!