Overview

Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off

One annual plan covers every course and certificate on Coursera. 40% off for a limited time.

Build the analytical skills that turn raw data into decisions leaders can act on. In this course, you will move through a complete decision-intelligence workflow — from exploring and summarizing data to running rigorous statistical tests, building production-ready predictive models, and communicating results to non-technical stakeholders. You will learn to generate descriptive statistics and visual summaries that reveal data quality issues before they distort your analysis. You will design and execute hypothesis tests, interpret p-values in business terms, and balance Type I and Type II error trade-offs with confidence. In the modeling track, you will build and cross-validate classification models using scikit-learn, handle class imbalance with techniques like SMOTE and class weights, and apply feature-selection methods — including RFE and LASSO — to balance accuracy with interpretability. The course culminates in an end-to-end customer lifetime value prediction project that integrates every skill into a portfolio-ready deliverable. Whether you are moving into a data analyst, business intelligence, or machine learning role, this course gives you the technical depth and communication skills to stand out.

Syllabus

Confidence-Interval Estimation - Foundation

Apply confidence-interval estimation to compare conversion rates across segments and present the statistical significance.

Type I/II Error Trade-offs - Core Application

Evaluate Type I/II error trade-offs for a proposed test and recommend appropriate alpha and beta thresholds.

Two-Sample t-Tests & Power Analysis - Integration & Assessment

Conduct a two-sample t-test in Python/R, interpret p-values, translate outcomes into plain-language business recommendations, and analyze test power under varying sample sizes.

Multiple Linear Regression - Foundation

Build and diagnose multiple linear regression models with proper statistical validation and remediation techniques.

Classification Methods - Core Application

Apply advanced classification methods including gradient boosting and logistic regression while handling class imbalance for optimal performance.

Model Evaluation & Selection - Integration & Assessment

Evaluate and remediate class imbalance using SMOTE while documenting performance impact on F1-score for comprehensive model validation.

Random Forest Model Building - Foundation

Build cross-validated random forest models that achieve business-defined accuracy targets

Model Drift Evaluation - Core Application

Evaluate and monitor model drift using statistical metrics to ensure long-term reliability

Cross-Validation Pipelines - Integration

Implement standardized cross-validation pipelines for multiple supervised algorithms and compare performance metrics

Feature Selection Methods - Assessment

Assess feature selection techniques to balance model accuracy with interpretability

Project: Statistical Thinking & Predictive Modeling

You will build a complete customer lifetime value (CLV) prediction pipeline for an e-commerce company. Starting from raw transaction data, you will conduct exploratory data analysis, execute a hypothesis test comparing customer segments, build and cross-validate a classification model, apply feature selection to balance accuracy and interpretability, and deliver an executive summary memo with actionable marketing recommendations. The project integrates data summarization, statistical inference, classification modeling, and supervised learning into a single end-to-end analytical workflow.