Week 5 drew a trendline by eye. This week it becomes a numeric, multi-variable model you fit and then judge. Two models on the same Summit Gear customers: OLS forecasts how long an account takes to pay, and logistic regression estimates the probability an account defaults. The stance is AI-first: the AI writes the Python; your job is to specify the model, validate it, and read it in accounting terms.
Lab 8 and Homework 8 are posted below. Check Canvas for the due dates.
Self-paced deck. Correlation, OLS (fit, residuals, RMSE, R², reading a coefficient), the multicollinearity caution diagnosed with a correlation matrix, dummy variables and the base case, then logistic regression end to end. Every slide carries an in-deck AI tutor and a self-check.
Hands-on trainers you run in the browser: correlation, the OLS trainer, the Excel path, the base-case/dummy-variable trainer, logistic regression, and the feature-selection visualizer. Each one checks your work and carries the AI tutor.
Specify both models, let the AI fit them, and read the output. You write the spec, validate the fit, and complete a notebook that checks your work as you go, then hand in an AI-generated HTML explainer you assess. An optional walkthrough shows the OLS done in Excel.
Twenty multiple-choice questions on reading model output a colleague or an AI hands you: an OLS coefficient table, a correlation matrix, a logistic odds ratio, the look-ahead trap, and a confusion matrix. Auto-graded.
2,200 Summit Gear B2B customer accounts, the dataset behind every model this week. A pre-loaded Excel workbook, a SQLite copy, and a data dictionary are linked from the lab and homework.
The starting notebook for the lab: load the data, let the AI write the fits, fill three small helper functions, and run the final check. Also downloadable to work locally with the Claude extension.
The class AI is built into every page here. Open the Ask AI panel on the lecture, the launchpad, the lab, or any trainer. It can read what you have typed and nudge you toward the fix without handing you the answer. Use it while you work and tell me how it goes.