Inspiration
Aiming to benefit those within our community, we collaborated with a local small company, Sun Tropics, to determine what problem it and other businesses needed help with. Sun Tropics shared with us their struggle of handling unique POs (Purchase Orders) PDFs from different customers and converting them to a standardized format in an Excel spreadsheet, which is crucial for them to fulfill the orders. Their current approach is mundane and tedious, requiring employees to manually process POs due to each customer submitting unique formatting. Upon hearing of this struggle, we arrived at our solution: OrderMatic!
What it does
OrderMatic is an innovative web application that implements what we’re coining as dynamic automated formatting. It’s simple: all the company has to do is upload a customer’s PO PDF along with the template Excel spreadsheet in the company’s standardized format. Our program then analyzes the data from the PO document and creates a new downloadable spreadsheet populated with the data from the original PO. OrderMatic handles variations within PDF formatting with ease, freeing employees to dedicate their efforts towards more significant tasks.
How we built it
We created a user-friendly frontend using ReactJS and connected it to a Python backend using Flask. For secure authentication, we implemented Firebase to keep track of users. To analyze the data from the PO PDF, we send it to the OpenAI API, from which our program creates a new spreadsheet in the desired format.
Challenges we ran into
We ran into a multitude of challenges, from CORs errors when connecting the backend and frontend to difficulty passing the PDFs and Excel template to the AI model. The main challenge we faced was getting the backend to successfully create a new spreadsheet in the correct format. Many modules we attempted to use to gather data from the spreadsheets were poorly maintained, resulting in a lot of errors. Thankfully, we pushed through and created an application we’re proud of!
Accomplishments that we're proud of
We are proud to have been able to get the AI model to accurately glean data from the PO documents and populate it on a new spreadsheet. A lot of times, the AI model was unsuccessful in its calculations or captured the wrong numbers, but after a lot of tinkering with the prompting, we finally got a reliable and accurate output. We are also pleased with the user-friendly interface of our website, which required the combined efforts of our team to design.
What we learned
Throughout this project, we learned how to download, edit, and save various forms of files programmatically from PDFs and Excel sheets to PNGs and CSVs. We also recognize the importance of having good time management. Even when our program was not fully finished yet, we delegated people to work on the presentation beforehand.
What's next for OrderMatic
Although this program was initially created for Sun Tropics, we truly believe it can be applied to help many other small companies with their POs as well as other types of documents. In the future, we plan to reach out to more companies and offer assistance as well as publish our website on a public domain.
Log in or sign up for Devpost to join the conversation.