Mijaaz is a cutting-edge music recommendation Android application built with Flutter. It leverages a custom-trained Artificial Neural Network (ANN) to provide personalized song suggestions for Pakistani and Indian music listeners.
| Splash Screen | Craft Your Vibe | Recommended | Spotify Integration |
|---|---|---|---|
 |
 |
 |
 |
Unlike most recommendation systems that rely on heavy cloud APIs, Mijaaz works 100% offline.
-
On-Device Inference: The ANN model is integrated directly into the app using TensorFlow Lite.
-
Smart Loading: During the Splash Screen, the app pre-loads the ANN model into the system RAM. This ensures that once the app opens, predictions are instantaneous.
-
No Internet Required: Since the model and the processed dataset (Spotify tracks optimized for Desi music) are built-in, users get recommendations anywhere, anytime.
-
Custom ANN Model: Trained on a Spotify dataset, specifically filtered for South Asian (Pak/Ind) musical characteristics.
-
Mood & Energy Analysis: Predicts the perfect track based on user-defined energy levels, danceability, and mood and some others.
-
Privacy Centric: No data leaves your device. All computations happen locally.
-
High Performance: Optimized RAM management during the splash screen for a lag-free experience.
Video is Also Given In the Repo (https://github.com/MubashirShafique/Mijaaz-App-Showcase/releases/tag/video_off_App)
-
Mobile Framework: Flutter (Dart)
-
Machine Learning: TensorFlow Lite (TFLite), Keras, Python
-
Data Processing: Scikit-learn (StandardScaler), Pandas
-
Dataset: Spotify Tracks (Curated for Pakistan & India)
The core ANN architecture and the Flutter source code are kept in a Private Repository for intellectual property protection.