https://github.com/belpa03/mobile-device-usage-user-behavior
Deep learning-based analysis of mobile device usage and user behavior using ANN. Includes data cleaning, EDA, feature engineering, model training, and infographic visualization.
https://github.com/belpa03/mobile-device-usage-user-behavior
ann deep-learning eda keras phyton tensorflow user-behavior
Last synced: 3 months ago
JSON representation
Deep learning-based analysis of mobile device usage and user behavior using ANN. Includes data cleaning, EDA, feature engineering, model training, and infographic visualization.
- Host: GitHub
- URL: https://github.com/belpa03/mobile-device-usage-user-behavior
- Owner: belpa03
- Created: 2025-10-05T08:35:56.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2025-10-05T08:55:29.000Z (3 months ago)
- Last Synced: 2025-10-05T10:27:33.293Z (3 months ago)
- Topics: ann, deep-learning, eda, keras, phyton, tensorflow, user-behavior
- Language: Python
- Homepage:
- Size: 762 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# 📱 Mobile Device Usage & User Behavior — ANN Deep Learning Project
---
## 🔍 Dataset Overview
**Source:** [Kaggle - Mobile Device Usage & User Behavior Dataset](https://www.kaggle.com/code/momenayman9/mobile-device-usage-and-user-behavio/input)
| Variable | Description |
|----------|-------------|
| User ID | Unique identifier for each user |
| Device Model | Smartphone model used |
| Operating System | Device OS (iOS or Android) |
| App Usage Time (min/day) | Average daily app usage in minutes |
| Screen On Time (hours/day) | Average daily screen-on duration |
| Battery Drain (mAh/day) | Daily battery consumption |
| Number of Apps Installed | Total installed apps |
| Data Usage (MB/day) | Daily data usage in MB |
| Age | User age |
| Gender | Male / Female |
| User Behavior Class | Behavior classification (1–5, light to extreme) |
---
## 🔁 Project Workflow
### 1. Data Ingestion & Initial Check
- Load dataset and inspect schema.
- Check for missing values and inconsistent types.
- No missing values found → ready for preprocessing.
### 2. Exploratory Data Analysis (EDA)
- Analyze distributions of User Behavior Class, Age, Gender, OS.
- Visualize correlations between App Usage Time, Screen On Time, Battery Drain, Data Usage.
- Insights: heavy usage correlates with high battery consumption; Class 2 dominates.
### 3. Data Cleaning & Preprocessing
- Encode categorical variables.
- Standardize numeric variables for ANN modeling.
- Dataset split into 80% training / 20% testing.
### 4. Feature Engineering
- Create derived metrics (e.g., usage per app, battery per hour).
- Correlation analysis to select features for ANN.
### 5. ANN Model Training
- Train ANN with training set, validation split 0.33, 100 epochs.
- Monitor accuracy and loss to ensure proper convergence.
### 6. Model Evaluation
- Evaluate test accuracy and analyze confusion matrix.
- Class 1 shows highest correct-classification rate (73.8%).
### 7. Predictions
- Generate predicted classes for test set.
- Evaluate probability distributions across classes.
### 8. Model Visualization
- Infographic charts for training accuracy, loss, and confusion matrix.
- Heatmaps and scatterplots for feature correlations.
### 9. Reporting & Business Insights
- Infographic + one-page summary for stakeholders.
- Provides actionable insights for app retention, device optimization, and personalization strategies.
---
## 💡 Conclusions
- Device usage intensity directly impacts battery consumption and app interaction.
- Behavior Class 2 dominates the population, while Class 1 is least represented.
- ANN achieved **76.62% accuracy**, showing good predictive performance.
- Confusion matrix shows Class 1 was most accurately predicted (73.8%).
- Misclassifications mainly occur between adjacent classes (2 and 3), suggesting overlapping behavioral traits.
- ANN effectively models nonlinear relationships, supporting segmentation and targeted interventions for different user behavior classes.
---
## 🖼️ Infographic Notes (`User Behavior.png`)
