Low-income housing is an issue that has faced the American people for decades, and it continues to plague our country to this day. Local and state governments have been working endlessly to put an end to homelessness and provide housing and vital resources to the less fortunate, but it hasn't been enough. As software engineers, we decided to use our skills for good and provide a platform to make it that much easier for those in need to quickly get back on their feet.
What it does
Our search engine scrapes websites like Craigslist to find furniture and housing based on provided budgetary information. Then, it uses artificial intelligence and natural language processing to filter through the results to provide custom recommendations based on the user's needs. To wrap it all up, the engine provides a simple search feature for those in need of work, powered by Indeed, creating a one stop shop for those who have been hit by hard times.
How we built it
We built Fetch by creating a fast, multi-threaded Flask API that communicates with the internet, then sends parsed text through the GCP AutoML NLP engine to create custom recommendations for users. The NLP engine is built to determine relevant postings on websites like Craigslist and Indeed, such as determining whether a posting is spam. This is followed up by an algorithm that uses dynamic programming to help users allocate their budget and resources in the most efficient way possible.
Challenges we ran into
It was difficult to come up with a technology in just 24 hours that would help solve a problem that governments have been spending millions of dollars to solve over the last few decades.
Accomplishments that we're proud of
None of us had ever needed to use dynamic programming outside of the classroom, so to see it being used in something as important as helping low-income families find housing.
What we learned
Surprisingly, rather than learning state-of-the-art technologies, we honed in our object-oriented programming skills and version control the most in order to create a clean and functioning product for this hackathon.
What's next for Fetch
We aim to make Fetch even faster, as well as add as much functionality as possible in order to create the true one-stop shop for those in need.