Decoding Regression Scores - Issue 147
How to read linear regression plots and understand the equation that drives analysis and predictions
Welcome to the Data Analysis Journal, a weekly newsletter about data science and analytics.
Last week I shared my recent publication in
newsletter about correlation analysis and linear regression. It was more of an introductory piece to the regression, and there is still so much more to cover on it.So today, it is Linear Regression Part 2, dedicated to my great analysts who are expected to work with ML and effectively understand the concept of regressions.
Linear regression is the most commonly used analysis to locate the relationship between values and run predictions. It has been widely adopted across product, growth, business, finance, and marketing by all levels of professionals.
What many people might not realize is that linear regression is a machine-learning model, so you have to treat it accordingly (considering data cleaning, feature engineering, addressing underfitting and overfitting, all of that). Reading its scatterplots is not conclusive. Many analysts apply regression without understanding what the scores and equation actually mean or how to read them. Let’s fix that today.
In this publication, I focus on:
Breaking down use cases across different regression types.
Examples of product and marketing questions and hypotheses that you can solve with regression analysis.
Explain the linear regression equation, what it means, and how to interpret the regression scores.
How to break down scatterplot slope using the regression equation.
Introduction to regressions and their use cases
Once again, regression shows how much one variable affects another and whether you can use the pattern of one variable to predict and estimate the behavior of another. Any type of regression doesn’t prove causation but rather gives us more clues and higher confidence that there’s a strong connection between variables, and how this connection will change if we increase or decrease variables.
Possible use cases for regression:
How many page views do we need to improve signups by at least twice?
Is showing more recommendations positively affecting trials?
How many more games/trials/clicks do users have to make to convert to Paid?
If we send 3x more notifications, how much will this increase DAU?
Keep reading with a 7-day free trial
Subscribe to Data Analysis Journal to keep reading this post and get 7 days of free access to the full post archives.