Decision Science and Strategic Modeling

Overview

The "Decision Science and Strategic Modeling" course equips learners with essential decision-making frameworks and methodologies. Students will explore the complexities of human decision-making, simulating bias and uncertainty while grasping decision theory fundamentals. Through structured decision tools like decision trees, influence diagrams, and programmatic modeling, learners will transition from theoretical predictions to practical decisions. The course emphasizes Bayesian thinking, sensitivity analysis, and the crucial connection between causal inference and actionable outcomes. Participants will also develop skills in communicating uncertainty, trade-offs, and recommendations effectively, culminating in a real-world pricing decision case study.

Syllabus

Course Introduction

Get an overview of the course structure, learning objectives, end-of-course project, and meet the course instructor.

Why Humans Struggle with Decisions

Explore why intuition leads to poor decisions, learn key cognitive biases, and see how structured decision models produce better, more consistent outcomes.

Simulating Bias and Uncertainty in Decision-Making

Learn to simulate how binary thinking and biases lead to errors in decision-making, using weather and other examples, and how to derive cost-aware, evidence-based strategies.

Decision Theory Fundamentals

Explore decision theory basics: expected value, expected utility, risk aversion, certainty equivalent, payoff matrices, and minimax regret for structured decision-making under uncertainty.

Computing Expected Utility and Comparing Decisions

Learn to build payoff matrices, estimate probabilities from market data, and compare decisions using expected value, utility, and minimax regret to guide real business choices.

Structuring Decisions with Trees and Influence Diagrams

Learn to structure messy business decisions with influence diagrams and decision trees, visualizing choices, uncertainties, payoffs, and grounding probabilities in real data.

Building and Solving Decision Models Programmatically

Transform business problems into Python decision models, using real data, influence diagrams, payoff functions, and trees to compare expected value and risk-sensitive choices.

From Predictions to Decisions

Learn to translate model predictions into actionable decisions using expected value, address cost asymmetry, and apply expected utility for robust, risk-aware choices.

Using Model Outputs in Decision Calculations

Learn to use model outputs—like default probability and loss severity—in formulas to maximize profit and manage risk when making business approval decisions.

Bayesian Thinking for Decision-Making

Learn Bayesian thinking: use priors, update beliefs with new evidence, and convert updated probabilities into actionable business decisions for better forecasting.

Updating Decisions with Bayesian Methods in Python

Learn to sequentially update business decisions and demand forecasts in Python using Bayesian methods, combining priors and data for actionable, confidence-based recommendations.

From Causal Inference to Actionable Impact

Learn to correct for selection bias using propensity scores and IPW, quantify causal effects with uncertainty, and translate them into actionable, financially-driven ROI decisions.

Using Causal Estimates in Cost-Benefit and ROI Analysis

Learn to correct naive program impact estimates with causal inference methods, then integrate reliable effect sizes into cost-benefit and ROI analysis with uncertainty bounds.

Decision Sensitivity and Scenarios

Learn to stress-test decisions by identifying key input drivers, using scenario analysis, and calculating break-even points and margin of safety to improve model robustness.

Running Sensitivity and Scenario Analysis in Python

Learn to build parameterized NPV models, run sensitivity and scenario analyses, create tornado diagrams, and solve for break-even points in Python with real-world examples.

Robust Decision-Making Under Uncertainty

Learn to make robust decisions by modeling uncertainty with distributions, understanding correlated risks, and ensuring defensible outcomes across varied scenarios.

Stress-Testing Decisions with Simulation and Uncertainty

Learn to use Monte Carlo simulation to model uncertainty, stress-test business decisions, and interpret outcome probabilities, moving beyond single-point estimates.

Communicating Uncertainty, Tradeoffs, and Recommendations

Learn to structure recommendations with BLUF, communicate uncertainty and tradeoffs clearly, use visuals effectively, and distinguish statistical from scenario-based uncertainty.

Presenting Decision Analyses with Visuals and Narratives

Learn to present decision analyses using purpose-built visuals and a BLUF memo, focusing on clear recommendations, key risks, and actionable insights for stakeholders.

Course Conclusion

Conclude the course by reviewing all learning objectives covered and reflecting on key takeaways from the course journey.

Price it Right — Nimbus Streaming Pricing Decision

Built a Bayesian decision model using IPW, Monte Carlo simulation, and decision theory to recommend the optimal pricing strategy for a streaming organization.