HappinessML: Predictive Analysis of Global Well-Being

Inspiration

My project, "HappinessML: Predictive Analysis of Global Well-Being," was inspired by the desire to understand the factors that contribute to the happiness of people around the world. By leveraging the rich dataset from the World Happiness Report, I aimed to explore how various socio-economic and personal factors influence happiness. My inspiration stemmed from a combination of academic curiosity and a genuine interest in contributing to global well-being research.

What it does

The project analyzes the World Happiness Report data from 2015-2019, identifying key factors that influence happiness in different countries. Using machine learning models, it predicts happiness scores based on these factors. The models we developed include Linear Regression, Ridge Regression, and Polynomial Ridge Regression, each providing unique insights into the data.

How we built it

I built the project using the following steps:

1. Data Collection: Sourced the World Happiness Report data from 2015-2019.
2. Data Cleaning and Standardization: Standardized column names, filtered relevant columns, and dealt with missing values.
3. Exploratory Data Analysis (EDA): Performed EDA to understand data ranges, types, and central tendencies.
4. Feature Engineering: Identified and engineered key features that influence happiness, such as GDP per capita, social support, life expectancy, freedom, trust in government, and generosity.
5. Model Development: We developed three models—Linear Regression, Ridge Regression, and Polynomial Ridge Regression.
6. Model Training and Evaluation: We trained the models on the data from 2015-2018 and evaluated their performance on the 2019 data.
7. Visualization: We visualized model performance using scatter plots and residual plots and analyzed feature importance.

Challenges we ran into

1. Data Inconsistencies: Handling inconsistent entries across different years was challenging.
2. Missing Values: Dealing with missing values required careful imputation to ensure data integrity.
3. Multicollinearity: Managing multicollinearity among predictor variables like GDP and health indicators required the use of advanced regression techniques.
4. Model Selection: Choosing the best model to capture the complexity of the data and making accurate predictions was a significant challenge.

Accomplishments that we're proud of

1. Successful Data Cleaning and Standardization: We effectively cleaned and standardized the dataset, making it suitable for analysis.
2. Model Development and Comparison: We developed and compared three robust models, each offering unique insights.
3. Feature Importance Analysis: We successfully identified and analyzed the key factors influencing happiness.
4. Custom Happiness Score Formula: We developed a custom formula for predicting happiness scores based on our model insights and correlations.

What we learned

1. Importance of Data Preprocessing: Proper data cleaning and standardization are crucial for accurate model predictions.
2. Model Selection: Different models have their strengths and weaknesses, and selecting the right model depends on the data characteristics and the problem at hand.
3. Feature Engineering: Identifying and engineering the right features is critical for model performance.
4. Visualization: Visualizing model performance and feature importance helps in better understanding the results and communicating them effectively.

What's next for HappinessML: Predictive Analysis of Global Well-Being

1. Enhance Models: Explore more advanced models and techniques to improve prediction accuracy.
2. Expand Data Sources: Incorporate additional data sources to enrich the analysis and capture more factors influencing happiness.
3. Real-Time Predictions: Develop a real-time prediction tool for policymakers and researchers to assess the impact of various factors on happiness.
4. Interactive Visualizations: Create interactive visualizations to make the insights more accessible and engaging for a broader audience. By continuing to build on this project, I aim to contribute further to the understanding and improvement of global well-being.

Built With

• jupyter
• kaggle
• matplotlib
• numpy
• python
• scikit-learn
• seaborn
• tpot