CS467 Capstone Project - Alexander Dembo

So we’re about halfway through the development phase of our project and things are looking pretty good. We’re definitely on track for completing our project with the goals we set out with. This past week, my project members were able to nicely work through a large chunk of the front-end design and back-end integration. We’re at the point where we just need our database to interact correctly with our app, and we’ll be pretty much done. I spent most of the week transferring over the DB to OSU’s server and adjusting the DDL accordingly, since it was formatted for a different service. After that, I crafted a DML that would generate selects and inserts based off the front-end design. This was done using ChatGPT’s assistance for what types of joins to use to get correct data selection. I also used ChatGPT to assist with rewriting my db-connector since we’re switching from Google cloud services MySQL to MariaDB MySQL. The only actual change that was needed was port, hostname, and login, it essentially functioned the same. I will be making our Progress Report video this week and I will see if AI can help generate a script to hit the primary points and I’ll fill it in with more human language. We’ll see how that goes!

Now that the project has been fully assigned and my team and I have fleshed out the specs, roles, and requirements, I got to start working on the actual database. Since our project focused on AI tools and how to utilize them optimally, I started by asking ChatGPT how it would recommend setting up a database for a React Native Front-end and Flask back-end physical fitness tracking application. It started by recommending PostgreSQL, which I have never used before. I figured that learning a new DBMS would be useful as a tool under my belt. I set off to learn how to implement PostgreSQL and was first approached with the problem of how I was going to host and connect to my Database. Luckily, OSU provides us students with a coupon for Google Cloud Services, so I requested one and went over to their service setup to get started. With cost estimation, google stated that the cheapest available PostgreSQL instance would run at a cost of ~$3.38/day. With the $50 coupon OSU provided us, this would last 2 weeks, which was not long enough for us to reach the end of the term with a functioning project. I was unsure of whether I would be able to turn the instance on and off (you can) but it seemed to be a hassle overall. I talked to our grader and professor and decided that it would probably be best to run the database off the OSU MySQL server. Furthermore, I was already generating tables and an ERD in MySQL workbench, and so I was able to finish that work and forward-engineer a DDL. I used ChatGPT during this process to help with how information should be stored between tables. I specifically had trouble figuring out how we would track the entire workout history for all users, since we wouldn’t be making a new table for each user. After this trial and error portion, I got to work figuring out how to set up a Google cloud SQL server as a temporary point so we could interact with a database and write DML queries. After a 1.5 hour troubleshooting and reading documentation process, I got it up and running and was able to connect! ChatGPT was very unhelpful for this portion, providing very specific snippets of code that with little to no explanation about their functionality, and they didn’t fit the question at all. Finally, I was able to start adding our static data to our tables and am currently working on getting example data and talking with our back-end on how we’ll get the database connection setup through flask SQLAlchemy! Overall, a pretty eventful and productive 2 weeks!

Hi all! 

My name is Alex, and I’m taking CS467: Online Capstone Project, as my only summer class and final class at OSU for the Post-Baccalaureate Computer Science program. We just were assigned projects and my project focuses on utilizing AI to help write project plans, design, and code! It’s more of a research project than an app building project, so the app comes second to the research and report. The app we’ve (my team) decided to build is the personal fitness trainer app. This app essentially will allow users to create a user profile, fill out a survey, be given a workout routine, and then customize the routine. It will be filled with notifications and information for healthy and reliable workouts and personal fitness. We landed on this project specifically since all my group members (including me) seemed to have a medical background. It just seemed to fit really well for our group. Next steps are going to be creating a project backlog, assigning tasks, and chugging through towards a working app! I’ll be blogging here each day I work on the project about what prompts I’m using for AI and what kind of information I’m receiving. How helpful that information is and how difficult it can be to work with and utilize. Hopefully by the end, I’ll have the sanity and clarity to put together a nice research write-up about AI and it’s advantages (and definite disadvantages) as a tool for building a web app.