PigeonBot

Context

As we tackled this project, our first challenge was brainstorming an idea that both aligned with the theme and our skills. Initially, we were gearing toward the UI/UX track and Data Science track, however, we spent all of Day 1 simply brainstorming, signifying that we should change gears. Eventually, we landed on creating a chatbot that’ll provide real-time service alerts and information regarding MTA subway services. More than 8 million people travel using the MTA daily and more often than not, there may be service delays or updates regarding trains. Many people may not be informed about the updates and may not prepare to reach their destination on time. Our Chatbot will help people of any age group to be informed of real time MTA updates, which will allow them to navigate quickly and efficiently anywhere in New York City. Our game plan was to split the team in half based on our strengths in front-end and back-end, develop our sides, and then combine on the way.

Challenges

Our major challenge for this project was merging both branches, ensuring that the user interface is compatible with the back-end code and vice versa. Moreover, some of the systems we were using (Repl.it, VS Code, and CodeSandbox.io) weren't properly showing our code, making it difficult for us to narrow down the root of the problem. After several hours of constant editing, searching for solutions, and time running away, we decided it would be best to commit all the changes and additions to our repository.

Accomplishments

As our very first hackathon, we’re proud of what we achieved with such little time and with no background in AI development. Although the front end was familiar with web development, we polished and added to our skills in designing and coding basic functions in JavaScript. On the other hand, the back end wasn’t familiar with AI development and still committed several long hours to research, code, API search, and overall development. While we didn’t finish our project in time, we can proudly say that we built up both our technical and professional skills. This was our first time working with OpenAI and API keys. Through this hackathon, we learned about Langchain, which lets us use Open AI to read text from a .txt file and understand it to help answer any questions asked by the user.

Future

As for PigeonBot, we’d like to see it soar with us. We hope to push through the obstacles of merging and finish our project, ensuring it functions with the user interface. PigeonBot’s development was an enriching learning experience for us and our gateway to stepping into AI development.