Selected Projects in Data Science and Machine Learning

Flight Passenger Satisfaction Prediction (Master’s Final Project)

Goal: Analyze key factors affecting airline passenger satisfaction and build a predictive model.

Tools: Python, Pandas, Matplotlib, Seaborn, XGBoost, CatBoost

• Built 5 ML models and selected top 3 to create a high-performing ensemble.
• Achieved 95.36% accuracy using hyperparameter tuning.

🔗 View Notebook | View Presentation


Comparison of base model performances (XGBoost, CatBoost, Random Forest).

Hyperparameter tuning phase results per model.

Performance of the final ensemble model.

Top features influencing customer satisfaction.

Telco Customer Churn Prediction

Goal: Predict churn based on customer behavior and service attributes.

Tools: Python, Pandas, Seaborn, DecisionTreeClassifier

• Identified high-risk customer profiles using EDA and decision trees.
• Analyzed when and why users leave based on contract and support quality.

🔗 View Notebook


Most frequent churn reasons: customer support and better offers elsewhere.

Most churn occurs in the first 24–26 months of customer life cycle.

Longer contracts are linked with lower churn rates.

Iris Dataset Clustering with K-Means

Goal: Apply clustering to discover natural groupings in flower data.

Tools: Python, K-Means, PCA, Elbow Method

• Applied unsupervised learning and compared results to known species.

🔗 View on GitHub


Cluster visualization using PCA reduction.

Tourism Data Visualization (2020)

Goal: Analyze global tourism trends during the COVID-19 pandemic.

Tools: Tableau

• Created interactive dashboards to show regional impacts and travel changes.

🔗 View Tableau Dashboard


Tableau dashboard showing top destinations, spenders, and losses.