🎓 MSc Data Science, Universiti Malaya
🌏 Based in Malaysia
📊 Data science, machine learning, and analytics for real-world decision making

With a background in biotechnology and scientific research, I apply data science methods to analyze complex systems and support data-driven decisions.

My work spans time-series modeling, machine learning, and data analytics, combining structured analysis with domain knowledge. I focus on building reproducible workflows, interpretable models, and practical analytics solutions.

• Languages & Tools: Python, R, SQL
• Libraries: Pandas, NumPy, Scikit-learn, Statsmodels, Matplotlib, Seaborn
• Data & BI Tools: Power BI, Excel, Jupyter Notebook, MySQL Workbench
• Tools & Platforms: Git, VS Code
• Techniques: Data Cleaning, Feature Engineering, Machine Learning (Classification), Time Series Forecasting (ARIMA, ETS), Econometric Modeling (ARDL, VAR), SQL Window Functions, Data Visualization
• Other: Agile, Google-Certified Project Management, Scientific Documentation

🔹 Fuel Price Pass-Through & Inflation in Malaysia

Research project analyzing how global oil prices and exchange rates influence Malaysian fuel prices and transport inflation. Built multi-source dataset integrating oil prices, FX rates, fuel prices, and CPI Developed time-series forecasting models (ARIMA, ETS) Estimated pass-through effects using ARDL and VAR models Evaluated models using RMSE, MAE, and MAPE

🔹 Logistics Inventory Data Analysis (SQL + PowerBI)

SQL and Power BI analysis of shipment lead times, delay rates, inventory days, and SKU performance for a retail logistics context.

🔹 Recipe Site Traffic Prediction (Machine Learning + KPI)

Classified high-traffic recipes using Logistic Regression, Decision Tree, and Random Forest. Defined a business KPI — High Traffic Conversion Rate (HTCR) — to align model precision with strategy. → Best Model: Logistic Regression (Precision = 0.88, HTCR = 7.13)

🔹 Telecom Customer Churn Analysis

Predictive model to identify customers at risk of churn using billing and usage patterns.
F1 Score: 0.85 | Key tools: Python, Random Forest, Seaborn_

🔹 Insurance Claim Outcome Modeling

Built classifiers to predict insurance claims and explored risk segmentation.
Accuracy > 75% | SMOTE for class balancing_

🔹 Netflix Content Trends

EDA and visualization of global trends across genres, ratings, and durations.
Clear dashboards to support content strategy decisions_

🔹 Penguin Clustering (PCA + K-Means)

Unsupervised learning project to classify species based on biometric traits.

🔹 Crop Yield Prediction (Regression)

Modeled yield based on environmental factors to support precision agriculture.

🧠 More projects available in the Repositories

• Dashboard deployment with Streamlit
• Power BI data storytelling
• Model interpretability (feature importance, SHAP)
• Data pipelines and workflow automation

• 📧 Email: yievia@gmail.com
• 💼 LinkedIn: View Here
• 🗂️ DataCamp Portfolio: View Here