Inspiration
I was inspired to build this project as I have learnt python in university up to building classes, but I wanted to see what else I could learn in this short time frame about building an app and chatbot and putting it on a website. A personal hobby is volunteering at beach cleans and I am also interested in risk and linear regression so I wanted to integrate these into a chatbot.
What it does
It is a simple chatbot that helps users understand beach litter trends and predicts future risks using a SLRM, a basic risk model based on the Clean Coast Index, and aggregates from the Marine Conservation Society data. Users can ask questions about the data, risks, predictions and get recommendations for action based on these.
How I built it
I started with downloading the dataset, cleaning the data and choosing aggregates from exploratory data analysis. i then built a basic risk model based on the Clean Coast Index and mitigation and monitoring factors. Then I used a simple linear regression model to predict future trends and risks. I then built my chatbot functions with OpenAI, and used Flask to build the app and local website design. Finally, I used Render to upload the chatbot to a public website.
Challenges I ran into
I had to test it many times and change my prompt as I was updating which was difficult to know which prompts it would understand. I also had to build a function so that the AI could read my data, and use function calls in the chatbot so that the AI could use my functions which took a lot of trial and error.
Accomplishments that I'm proud of
These challenges taught me prompt engineering and how to implement AI to read data which I had not learnt previously. I accomplished building my first project with an AI chatbot and analysing data.
What we learned
I learnt how to build a risk model, make predictions with a simple linear regression model in python with new packages I had not used before, and how to implement these with an AI chatbot. I also learned how to build an app with Flask and put my chatbot on a public website with Render.
What's next for Beach Litter AI Chatbot
In the future I would add location based data analysis so the chatbot could provide information about areas with higher risk, add images and graphs to improve user experience, and expand to more environmental datasets.
Log in or sign up for Devpost to join the conversation.