Error and Uncertainty in Google Sheets

Learn to distinguish real differences from random noise, and explore psychological crutches we use that interfere with our rational decision making.

You rely on predictions every day: you might check the weather app before choosing your outfit or peek at the traffic before starting your commute. Perhaps you are responsible for setting your organization’s strategy in the future. Do you find yourself wondering how accurate predictions are, how you can see into the future, and why the weatherman always seems to be wrong? In our Error and Uncertainty course, you’ll make some predictions yourself, learn to distinguish real differences from random noise, and explore psychological crutches we use that interfere with our rational decision making. You will uncover patterns in Seattle crime data, predict students’ final grades, prevent Nashville traffic accidents, and determine whether a bakery’s menu needs to change. Join us! We’re certain you’ll enjoy learning about error and uncertainty.

• Defining error, uncertainty, and risk
• The first chapter presents common terminology, introduces methods for determining significant differences between groups, and outlines the kinds of error and uncertainty involved. We will specifically look at Seattle crime data and evaluate crime rate differences between precincts and neighborhoods. This chapter will equip learners to identify threats to the validity and accuracy of their conclusions.
• Making accurate predictions
• The second chapter outlines both rudimentary (e.g., moving average, seasonal average, yearly average) and more complicated methods (e.g., linear regression) for making predictions and outlines the kinds of error and uncertainty involved. We will specifically look at anonymized student grades data and evaluate the accuracy of our predictions for given students. Throughout the chapter, we will identify threats to the validity and accuracy of our predictions.
• Poking holes in predictions
• Chapter 3 encourages learners to test the assumptions of their predictions using data on car crashes. Specifically, they will determine how to allocate resources to reduce injuries and fatalities from auto accidents. Learners will discuss the impact of outliers in prediction accuracy, evaluate the importance of normally distributed data in making predictions, employ consequence-likelihood matrices in risk management, and adapt psychological heuristics to discussions of numerical uncertainty and risk.
• Case study: Should you change your bakery's menu?
• The final chapter integrates all the previous lessons into a constructed-world scenario. Learners are tasked with updating the menu at their small business: the Risky Business Bakery. They need to figure out whether to add or drop menu items based on whether there are significant differences in sales by baked good; whether their predicted sales figures from their accountant are accurate.