Data Analysis for Managers

What you'll learn

• Frame analytical work around decisions using a decision brief, and question any analysis by asking compared with what, how big, how uncertain and what else could explain it.
• Describe data with appropriate summaries and charts, using medians, percentiles and boxplots for skewed business data, and design dashboards that separate signal from noise.
• Apply conditional probability, base rates and Bayes' rule to screening and alarm problems, recognize correlated risks and score probability forecasts for calibration.
• Identify sampling frame, selection, non-response and response bias, and calculate standard errors, confidence intervals and required sample sizes for proportions and means.
• Design and analyze randomized A/B experiments using two-proportion tests, interpret p-values correctly, and plan sample sizes for adequate power while avoiding peeking.
• Compare groups with Welch's t-test and paired or difference-in-differences designs, and control false positives from multiple comparisons with Bonferroni and Benjamini-Hochberg procedures.
• Distinguish correlation from causation by identifying confounding, reverse causation, Simpson's paradox and survivorship bias, and choose designs that support causal conclusions.
• Fit and interpret simple and multiple regression models, including slopes, R-squared, prediction intervals, omitted variable bias and collinearity.
• Use dummy variables, interactions, logarithms and logistic regression in business models, run diagnostics and validate models on held-out data.
• Forecast seasonal time series with benchmarks, exponential smoothing and seasonal indices, measure accuracy with MAE, MAPE and bias, and separate forecasts from targets and plans.
• Structure decisions under uncertainty with payoff tables and decision trees, and calculate expected value, EVPI and EVSI, supported by sensitivity analysis and simulation.
• Integrate the course methods into an evidence-based recommendation for an organization, stating uncertainty, tracking measures and responsible use of customer data.