Week 4: Design Phase (Software Methodology, Tool Setup & Approval)

This week was all about locking in the tools and technologies I’ll be using for the project and getting my development environment fully ready. After receiving approval on the project’s methodology, I shifted focus toward setting up everything I need to start building and training the system.

First, I installed Visual Studio Code, which is now my main workspace for writing and organizing Python scripts. It’s lightweight but powerful and helps keep everything in one place—from backend scripts to code related to model interaction. For the mobile side of the project, I set up Android Studio, which includes an emulator that lets me test the app directly on a virtual Android device. This is useful during development because I can quickly preview how everything looks and behaves without needing a physical phone every time.

For the machine learning aspect, I’ve been using Roboflow to handle all dataset preparation and annotation. It’s a web-based platform where I upload my images, label them using bounding boxes, and then export the dataset in formats compatible with YOLOv8. Once the dataset is ready, I use Google Colab to train the model. Colab’s cloud-based environment provides access to GPUs, so I can run training jobs more efficiently without worrying about the limitations of my local machine.

Another critical component of my project is Firebase, which I’m using as the backend for the app. The app is being developed in a way that allows it to communicate with Firebase to both store and retrieve data in real time. For example, once a ball is detected and analyzed, the relevant results—such as speed or trajectory information—can be uploaded to Firebase. This data can then be pulled and displayed on the user’s app interface almost instantly. Firebase helps simplify this entire flow by offering real-time database support and secure cloud storage, making it ideal for this kind of mobile-integrated application.

Everything is starting to come together: Visual Studio Code for managing the core scripts, Android Studio for building and testing the app, Roboflow and YOLOv8 for dataset handling and training, Google Colab for running the model, and Firebase for managing real-time app data. With the environment now fully set up, I can move forward with the actual development process—writing logic, building out features, and eventually connecting the app to the model results.