Study Club

Framing the Problem

What is the problem you are trying to solve? Who does it affect?

One of the biggest issues faced by students, and from our personal experience, is finding adequate study tools and being engaged while studying. Many students often find studying boring and not engaging enough due to many factors, an example being not knowing strong study skills, or unable to focus over extended periods. This, often, would cause students to lose interest in studying for an exam or quiz and get lower marks than their potential. We seeked to create a platform that uses AI to make  studying easy to track and engaging, creating an incentive to keep track of their progress.

Idea Explanation

What is your idea? How does it fix the problem?

Our idea is to combine multiple ways to study in one place. Instead of having to go to several sites for different methods or ways a user might prefer to study they can do it all in one place. Additionally it makes it easy to collaborate directly with others virtually compared to some sites where the collaboration is geared more towards in person usage.

Implementation

How do all the pieces fit together? Does your frontend make requests to your backend? Where does your database fit in?

In order to implement this, we decided to use Figma Make to create the website through a multitude of prompts. We first started by making the general design and creating the different tabs and UI for the website, without any of the intricate details and information of each tab, just to get a base. After we got the base we were looking for, we then started to slowly create specific and long prompts to create the features we wanted it to include, such as AI generated study guides, access to the internet and games using flashcards the user can create or from the internet. In order to do that, we created a supabase in github that is attached to the figmamake so that it can store the data from the internet. At first, we weren’t sure whether to use an API with data of specific topics, like physics or calculus,  or try to incorporate another AI into figmamake so it can generate questions and answers. Then, we used a service account API from Open AI so that figma make can access the web through chatgpt, create questions or study guides, store them in supabase, and allow the user to call on them whenever needed, such as in study games.

Challenges

What did you struggle with? How did you overcome it?

We struggled with the user interface and getting the API to work. For example buttons wouldn’t work, and scrolling features were a recurring issue. Also the general aesthetics were an issue. But fixing those issues, we fed FigmaMake reference photos, which helped get it more precise, as well as repeated emphasis on what we wanted to fix for the UI games and fun features. The connection to the web and document reading was another issue, we couldn’t get that to work using just Figma AI, and as this was our first time using AI generation and doing a project of this depth, we also struggled to figure out how to use API, github and databases. We overcame this through further research and uses OpenAI API guidelines to figure out how to use their API to allow figmamake to access the web to generate the products we wanted it to. 

Accomplishments

What did you learn? What did you accomplish?

We learned a lot about prompt engineering, how supabase and other databases interact with API keys and AI generation. We also learned a lot about the use of json files in reading documents and pdfs as well as different use cases for API’s, such as using OpenAI’s API for access to the web, or using other API libraries for specific information. The project came out a lot better than we hoped, working far more efficiently over the course of these 48 hours, exponentially after the incorporation of the OpenAI API.

Next Steps

What are the next steps for your project? How can you improve it?

The next step is to actually add our own personal touches. Since we were in a rush we heavily relied on AI for the production, and we want to learn more about the coding process that the AI took. We think we can improve it by adding more specific features and further prompt engineering the website and generate far more specific and indepth study guides and questions. Due to the time constraints, the generations made by this website are not that far from surface level information and study materials, so we hope to improve it by further allowing it to dive deeper into the information and create more thorough study materials for students.