Building My First AI Project: What Went Right and What Went Wrong

After learning different concepts in Computer Science and AI, I wanted to build something that was more than just another college assignment. That is how I started working on my first major project — AI Civic Issue Mapper. The idea was simple: what if people could report everyday civic problems like garbage dumping, potholes, water leakage, or broken streetlights using photos and location information, and the complaints could then be tracked properly? That simple idea turned into a much bigger project than I initially expected. The main goal of the project was to create a platform where citizens could report civic issues and track what happened to their complaints. A user can report an issue by providing information such as: 📸 A photo of the problem 📍 Location/GPS information 📝 Description ⚠️ Urgency level The system then helps organize the complaint, route it to the appropriate department, detect possible duplicate reports, and track its status. I also wanted the project to focus on transparency, so there is a public page where people can see complaint-related statistics without needing to log in. For the project, I worked with technologies that I was already learning as a Computer Science student: Python Flask MySQL HTML, CSS & JavaScript Leaflet.js for maps Google OAuth for login Brevo API for emails Render for deployment One of the biggest things I learned was that knowing individual technologies is very different from actually connecting them into one working application. The most satisfying part was seeing the project actually work. I was able to build features such as: User registration and login Google login Issue reporting with images and location Automatic department routing Duplicate complaint detection Admin dashboard Complaint status tracking Resolution proof uploads SLA tracking Public transparency page Complaint tracking without login Responsive pages for different screen sizes Eventually, I also deployed the application so that it could be accessed online. That moment felt very different from simply running a project on localhost. Of course, everything did not go smoothly. One of the biggest lessons was that building a project involves a lot more debugging than coding from tutorials. I had to deal with problems involving: Connecting the Flask backend with MySQL Managing environment variables Authentication Email delivery Deployment Database changes Handling errors properly Making pages responsive Testing different parts of the application Sometimes I would fix one problem and accidentally create another one somewhere else. There were also times when something worked perfectly on my local machine but behaved differently after deployment. That was probably one of the most frustrating parts. This project taught me something that tutorials cannot fully teach you. You don't really understand a technology until you have to use it to solve an actual problem. I learned how different parts of a software project depend on each other. For example, adding a simple feature could require changes to the frontend, Flask routes, database structure, validation, and sometimes deployment configuration. I also learned that building a project is not about writing thousands of lines of code. It is about understanding the problem, breaking it into smaller pieces, testing those pieces, and being patient when things don't work. Interestingly, the project is called AI Civic Issue Mapper, but the AI image-classification part is still something I want to improve. The current version focuses more on the complete civic-reporting system, while AI-based image classification is part of the future roadmap. I actually think this is an important lesson too. A project does not have to be "100% complete" before you can learn from it or share it. There is always another feature, another improvement, or another problem to solve. The project is still evolving. Some of the things I want to work on next include: 🤖 AI-based image classification 🗺️ A complete multi-complaint map 🌡️ Geo-spatial heatmaps Better testing Further improvements to the user experience I also want to continue improving my understanding of AI and eventually integrate what I am learning into projects like this. Building my first major project made me realize that learning Computer Science is not just about completing a syllabus. It is about taking what you learn and trying to build something with it. There were bugs, confusing errors, failed attempts, and plenty of moments where I had no idea what I was doing. But there was also something much more valuable — the feeling of finally seeing an idea turn into a working application. And honestly, I think that is one of the best ways to learn. This is only my first major project, and I still have a lot to learn. If you are also a student who is thinking about building your first project, my biggest advice would be: Start before you feel ready. You will learn a lot along the way. 🔗 Project: : https://github.com/Anushka190921/ai-civic-issue-mapper 🌐 Live Demo: : https://ai-civic-issue-mapper.onrender.com/