What sparked my interest in this project was my dad. He said he needed a tool to help him enter a lot of data into a spreadsheet for his insurance. He also said if there was a way to use ai to help him fill in some of the price columns, that would be a lifesaver because for some of the items he didn't know how to estimate a price for them. I felt like this would be a great project for me and I wanted to get a lot more familiar with dataframes so this was great.

Going into this project, I had more experience manipulating ai responses than I did with an excel sheet or dataframes ironically. I made a small project before where I was generating responses from ChatGPT based on some inputs so I felt like I could build on that knowledge in this project. I was excited to bridge the gaps in my knowledge however in gpt-3.5 and in dataframes.

The research for this project started off with watching many videos introducing me to streamlit and dataframes. I had experience with python but not much at this point so it was great to get some. I started researching how to setup streamlit normally and referenced their documentation so much especially when it came to integrating the ai components.

 

Once I was able to get streamlit setup, I then went onto learning how to create and edit dataframes. It took several videos and reading the documentation but I got the hang of being able to generate a dataframe and connect my own data to my application.

 

The most challenging part of this project was getting the ai output to be consistent when prompted but I did figure it out. The ai feature in this application allows users to fill in a spreadsheet, all of that information is sent to a chatgpt function, and it estimates a price for that item you gave to it and enters it into your spreadsheet.

 

So in this projects I really got to see the power of variables up close, Learned how to interact with chatgpt in a lot more nuanced ways, set up and deployed my streamlit application, and even added extra features like being able to download your database as a csv file when you're done.