World Happiness Report Data Analysis Presented By: Jacob - PowerPoint PPT Presentation

World Happiness Report Data Analysis Presented By: Jacob Bissonette, Kyle Costello, Zachary Mailloux, and Adrianna Staszewska

What are we looking at? - Happiness standings of different countries - Different factors that play important roles in a countries happiness rating - Happiness rankings of different regions of the world

What Does Our Data Include? - Results from 134 Countries between 2015 - 2019 - Dependent Variable: Happiness Score, Scale of 1 - 10 - Independent Variables: - Economic Score (GDP per Capita) - Health Score (Life Expectancy) - Freedom Score (ability to make own life choices) - Trust Score (higher score means greater trust/less corruption in gov’t) - Generosity Score

Cleaning the Data - Missing countries across data sets - Missing independent variables across data sets - Zeroed values for some independent variables in specific countries - Added extra variable called Regions for further break down

Approaching Our Problem Goals Approaches - Predict a happiness score for a country - Regression analysis - Determine trends in happiness scores with respect - Singular/Multiple/Tree to regions - Basic data visualization - Determine what variables have a larger deciding - Heat map factor in the happiness score

Initial Analysis

Regression Analysis

Regression in Excel Examined a Multiple Regression using the entire data set from 2015-2019 R 2 = 0.745 Independent Variables and Significance GDP P-Value = 5.81E-47 Life Expectancy P-Value = 8.66E-17 Freedom P-Value = 2.25E-23 Trust P-Value = 0.127080097 Generosity P-Value = 0.018402025

Choosing our predictors

Regression Tree Approach

Multiple Regression Model HappinessScore = 2.23853 + 1.34888*GDP + 1.49602*LE + 2.36287*F Multiple R 2 value for training data: 0.755 Multiple R 2 for testing data: 0.688

Comparison of Multiple Regression and Tree Methods Multiple Regression Tree Regression MSE 0.35 0.46 MAPE 9.33% 10.5% R 2 68.8% 58.4%

Conclusions - Most important factors when considering global happiness are: - GDP - Freedom - Life Expectancy - Using the Multiple Regression approach provided lower MSE values and higher R 2 values than the Tree Method approach - Determined that Europe, North America, and Oceania have some of the highest happiness scores in the world

Questions?