Geu-Pro2023/IndabaX_Hackathon_South-Sudan
Record type:
project
Creator:
Geu
Host:
IndabaX Hackathon SS
# IndabaX Hackathon South Sudan 2025
## Application of Machine Learning Techniques in Real-Time Data
A real-time weather classification web app built for the **IndabaX South Sudan 2025 - Intermediate Track Hackathon**, under the theme:
## Project Overview
This project is a submission by **DREAMERS** for the IndabaX South Sudan AI Hackathon. The competition challenged participants to build a weather condition classification model and **deploy it with a real-time user interface** to boost social impact through applied AI.
Our solution includes:
- A trained CNN model using EfficientNet/MobileNet on a labeled image dataset.
- A fully deployed FastAPI backend for real-time inference.
- A modern, responsive frontend built with HTML, Tailwind CSS, and vanilla JavaScript.
- A UI that allows users to upload an image and see predicted weather conditions instantly.
## 🌐 Live Demo
🔗 FastAPI: Click: dreamers-weather-classifica…
🔗 Real Website: Click: indaba-ml-app.netlify.app
## Model
- **Architecture:** Transfer learning with MobileNetV2
- **Accuracy on kaggle leaderboard:** Achieved over 98.67% test accuracy
- **Framework:** TensorFlow 2.16 (CPU)
## Accuracy we used to deploy the model: 98.67%
## Features
- Upload an image to classify the weather condition
- Supports categories like sunny, cloudy, rainy, foggy, etc.
- Real-time prediction using REST API
- Fully responsive design
- Weather-inspired color theme
## Team DREAMERS
We are a team of five final-year software Engineering students passionate about AI for social impact from African Leadership University (ALU), Rwanda
| DREAMERS | Program | Specialization | Contribution | School
|------------------------------|------------------------|------------------------|-------------------------------------------|------------------------------
| 1. Madol Abraham Kuol | Software Engineering | …