Week 14 model critique analyzing Week 7 hypothesis testing lab. Creates digital marketing agency business scenario testing display ads vs. standard ads for 50+ e-commerce clients. Identifies critical statistical issues: untested assumptions (normality, equal variance), temporal dependence violating independence, and multiple comparisons inflation. Provides 5 improved analyses with R code: comprehensive assumption testing (Shapiro-Wilk, F-test, Q-Q plots), autocorrelation checks for temporal patterns, Bonferroni/Holm multiple testing corrections, practical significance framework with confidence intervals and business decision matrix, and post-hoc power analysis with sensitivity curves. Examines 3 ethical concerns: selection bias from unverified randomization (justice principle), cherry-picking metrics and HARKing enabling misleading results (scientific integrity), and economic harm from Type II errors with $2.7M opportunity cost calculation (harm minimization). Demonstrates that rigorous data science requires both technical proficiency and ethical reflection.

Week 12 analysis examining temporal patterns in Ames housing market. Creates time series from Yr.Sold/Mo.Sold, aggregates to monthly median prices, detects linear trends with piecewise boom-bust-recovery phases, performs STL seasonal decomposition revealing $20K summer-winter swing, and illustrates autocorrelation with ACF/PACF showing trend, seasonality, and persistence.

Week 11 synthesis building complete multiple regression model with 7 predictors (size, quality, age, features). Includes full diagnostic analysis using 5 plots from class, identifies heteroscedasticity and right-skew issues, and interprets 3 key coefficients in detail. Demonstrates progression from exploratory analysis through model building to valid statistical inference.

Week 10 analysis using logistic regression to predict whether homes achieve premium quality ratings (8-10 out of 10). Models probability using living area, year built, garage size, and central air. Interprets coefficients as odds ratios, constructs confidence intervals using standard errors, and evaluates model performance with confusion matrix and accuracy metrics.

Week 9 revised analysis building multiple regression model with living area, quality rating, and central air variables. Includes comprehensive diagnostics using the 5 plots from class: residuals vs fitted, residuals vs X values, correlation heatmap, residual histogram, and Q-Q plot. Evaluates linearity, normality, homoscedasticity, and multicollinearity with severity assessments and confidence levels for each assumption.

Week 8 analysis using ANOVA to test house style effects on price and simple linear regression to model the relationship between living area and sale price. Includes assumption checking, effect sizes, diagnostic plots, and practical recommendations for buyers and sellers.

Week 7 analysis using both Neyman-Pearson and Fisher frameworks to test whether central air conditioning and recent construction affect home sale prices. Includes power analysis, sample size calculations, and detailed interpretations of statistical evidence.

Week 6 analysis examining relationships between original and derived variables. Calculates correlation coefficients, builds confidence intervals for population inference, and interprets statistical relationships in practical real estate contexts.

Week 5 analysis examining unclear data encodings, missing value ambiguities, and data quality problems. Investigates what happens when documentation is incomplete and demonstrates methods for detecting inconsistencies and defining outliers.

Week 4 analysis examining how conclusions change based on which random sample we collect. Creates multiple samples at different sizes (25%, 50%, 75%) to understand when findings are robust vs. sample-dependent. Demonstrates the relationship between sample size and reliability.

Week 3 analysis investigating group probabilities and anomaly detection in Ames housing data. Identifies rare building types, quality-neighborhood mismatches, and architectural combinations. Includes probability calculations, rarity classifications, and testable hypotheses.

Comprehensive data analysis of 2,930 Ames, Iowa home sales examining price patterns, neighborhood segmentation, quality-size relationships, and market stability during the 2008 financial crisis. Includes statistical summaries, visualizations, and actionable insights for buyers, sellers, and investors.