A Product Data Science–oriented telecom analytics project that transforms telecom usage data into actionable insights across customer overview, engagement, experience, and satisfaction workflows.
This project analyzes telecom customer behavior using Python, exploratory data analysis, feature engineering, clustering, and an interactive Streamlit dashboard.
It is designed to communicate the kind of workflow a Product Data Scientist would use to:
-
understand user behavior - profile customer engagement
-
explore experience-related metrics - segment users into meaningful groups
-
present insights through a clean interactive dashboard
-
The repository includes:
-
a Streamlit app for dashboard delivery
-
data files for cleaned and intermediate analysis outputs
-
notebooks for analysis work
-
Python scripts for cleaning and helper utilities
This project is positioned as a product analytics + behavioral segmentation case study in the telecom domain. Key product questions this project helps answer:
- Which users are most active and valuable?
- How does usage differ across apps like YouTube, Google, Netflix, Email, Gaming, and Social Media?
- Which users show high engagement versus low engagement?
- How can clustering help identify meaningful behavioral segments?
- How can telecom experience and satisfaction analysis be extended into retention and churn use cases?
- dataset quality checks
- missing value analysis
- handset manufacturer and handset type exploration
- user/session distribution
- total usage patterns across application categories
- sessions per user
- total duration per user
- total traffic per user
- top users by usage
- outlier detection
- KMeans clustering for engagement segmentation
- application-level usage comparison
- service quality indicators
- latency and reliability metrics
- friction analysis
- user experience scoring
- NPS-style metrics
- churn signals
- retention storytelling
- satisfaction driver analysis
Telecom_Analysis/
├── Data/
├── Notebooks/
├── scripts/
├── app.py
├── preview.png
├── requirements.txt
└── README.md
Core:
- Python
- Pandas, NumPy
- Scikit-learn
Visualization:
- Plotly
- Matplotlib
- Streamlit
Workflow:
- Jupyter Notebooks
- GitHub
- Virtual Environments (uv / pip)
Clone the repository
- git clone https://github.com/Denis0242/Telecom_Analysis.git
- cd Telecom_Analysis
Create and activate a virtual environment
- uv venv --python 3.11
Windows
- .venv\Scripts\activate
Mac/Linux
- source .venv/bin/activate
Install dependencies
- pip install -r requirements.txt
Run the dashboard
- streamlit run app.py
- loading the default dataset from Data/cleaned_data.csv
- optional upload of CSV or Excel telecom datasets
- dataset overview metrics
- missing value inspection
- handset analysis
- user behavior analysis
- application usage analysis
- engagement distributions
- outlier detection
- engagement clustering using KMeans
- User Overview Analysis
- User Engagement Analysis
- Experience Analysis
- Satisfaction Analysis
- cleaned_data.csv — main cleaned dataset used by the dashboard
- data.csv — original telecom dataset
- user_engagement.csv — engagement-focused data
- user_experience_metrics.csv — experience-related features
- Week1_challenge_data_source.xlsx — source dataset
This shows a full pipeline: raw data → cleaned data → analysis → dashboard.
- Exploratory Data Analysis
- profiling user behavior
- identifying missing data
- comparing usage across app categories
- Feature Aggregation
- session counts
- total duration
- total traffic
- app-level usage
- Behavioral Segmentation
- scaling and normalization
- KMeans clustering
- elbow method
- Dashboard Communication
- business-readable metrics
- visual storytelling
- interactive exploration
- translates raw telecom data into business insights
- applies product thinking to analytics
- builds interactive dashboards
- combines EDA + ML + storytelling
- demonstrates end-to-end workflow
Relevant Roles:
- Product Data Scientist
- Product Analyst
- Customer / Growth Analyst
- Behavioral Analytics
- churn prediction
- retention cohorts
- feature importance (SHAP)
- KPI layer (AARRR, North Star)
- A/B testing
- experience analytics
- Streamlit Cloud deployment
Denis Agyapong
Product Data Science / Data Analyst