Problem

We sought to interpret the potential relationship between CO2 emissions and GDP per capita across different countries. The project required comparing multiple visualizations and performing statistical analysis to test our hypothesis about how emissions scale with wealth.

Initial explorations revealed that the data were noisy and varied by region, making it hard to draw consistent conclusions from visuals alone. We needed a reproducible approach that could quantify the relationship and explain observed patterns across diverse economies.

Solution

We leveraged Python libraries such as NumPy, Pandas, and Matplotlib to clean, organize, and visualize the data, then conducted a linear regression analysis to test our hypothesis. This allowed us to quantify the relationship between CO2 emissions and GDP per capita and identify how it varied across countries and regions.

Our approach emphasized reproducibility and interpretability, documenting each step from data ingestion to model interpretation. We produced clear visualizations and a statistical summary to communicate insights to stakeholders and to support further policy discussions.

Conclusions

We came into this experimental analysis with the idea that a country's CO2 emissoins per capita were related to its GDP per capita. After performing statistical analysis on each data set, we have come out with several findings. First, both data sets were significantly skewed to the right, illustrating that the majority of both the coutries' CO2 emissions and its GDP are at lower values, with a few outliers at higher values. This shows us that most countries of the world have GDPs close to the global median value, and a few countries have much higher GDPs, with the same being true for each country's CO2 emissions.

Furthermore, having calculated a coefficient of correlation of 0.519, we can see that there is a positive linear relationship between GDP and CO2, meaning that as the GDP of a country increases, its CO2 emissions will increase, on average. From our scatterplot, we can also see that variability increases at higher GDP values. This tells us that countries with large GDPs (in other words, likely more developed, industrialized countries), have a large range of CO2 emissions. After performing a least squares linear regression, we calculated a coefficient of determination of 0.269, which is relatively low. This indicates the regression line we found was not a very good fit. The least squares regression line is the line which minimnizes the sum squared error of all observed data, so perhaps there would be a nonlinear model which would fit our data better.

From our results, we can see that there is a slight positive linear association; however, since the line we calculated had a low 𝑟2 , the relationship is not very strong. This could be due to the outliers, and the fact that our data sets were both not normally distributed.

Overall, the GDP of a country and its CO2 emissions per capita have a mild positive correlation. However, a low coefficient of determination indicates a linear model may not be the best fit for the relationship between the two data sets. In general, we can conclude that as GDP increases, CO2 can be expected to increase, on average, with the exception of a few outliers.

Jupyter Notebook: https://drive.google.com/file/d/1bf8LdEJGAR3OGA4eU1vSG2y0pCiCbZwq/view?usp=sharing

CIVENG 93 Engineering Data Analysis: CO2 vs GDP per Capita Regression