Native Plant Finder

Inspiration

KU has had a lot of efforts into introducing native plants into campus such as the removal of tulips, and we thought we could help increase awareness of invasive species by improving the public's perspective of native flowers.

What it does

Given a state and short description of a flower, our vector database uses multiple web sources to produce the best matches of native flora for the description.

How we built it

Using Beautiful Soup, we scraped multiple free web sources to obtain simple data and images of flowers. We then used Weaviate to create a vector database to store this data and Flask to host a server providing it to the end user in a simple webpage format.

Challenges we ran into

Finding the right web sources that provided locational data of plants that we could scrape
Finding a vector database to use for our search arguments
Formatting our data for user input and output

Accomplishments that we're proud of

"Implementing vector databases" - Michael
"Learning web scraping" - Sam
"Creating a frontend and working with Flask for the first time" - Blair
"Finding a solid web source and figuring out how to scrape it" - Dylan

What we learned

This was all of our first times using Flask and Docker, and we definitely plan on using both in the future
How to utilize vector databases for searches and relevancy
Google doesn't like web scraping
How to manipulate URLs for scraping