Data-focused professional with expertise in Python, SQL, Power BI, and machine learning. Experienced in data preparation and quality assurance, with applied projects spanning business intelligence, statistical analysis, forecasting, and predictive modeling. Combines analytical rigor with business understanding to deliver reliable insights, performance reporting, and data-driven solutions.
Transactions discounted above 20% were 15.79% of sales and carried negative profit (-$135,376), which is -47.27% of total profit.
18 of 74 product categories (24%) generated 81.26% of revenue, and review scores fell as delivery times passed about 14 days.
Daily forecasts failed (MAPE about 126%). Monthly SARIMA reached about 19% MAPE on a three-month test, which is moderate and thinly tested.
Month-to-month contracts, short tenure, fiber optic service and electronic-check payment stood out as churn risk factors; tuned LightGBM reached 0.810 recall.
In this dataset, revenue is concentrated in a few retailers and the top-selling categories are not always the most profitable.
Recorded rates were highest among men aged 75+, and male rates moved far more than female rates around the mid-1990s.
Coursework included Linear Algebra, Calculus, Quantum Mechanics, Computational Physics, Statistics and Statistical Physics, with practical MATLAB work on problem-solving and simulation.
IBM on Coursera
IBM on Coursera
Google on Coursera
Knowledge Accelerators on Coursera
Mostafizur Rahman