5  Conclusion

There are several key points we would highlight in conclusion. First is the dataset’s completeness changes a lot over different years. From 1990 to 2007, the data is mostly there (only 10% to 25% is missing). But from 2008, more data is missing. By 2011, most of the dataset is missing, making it unreliable for those years. This means we should be careful when using this data to look at trends, especially in recent years. Also we had variability across Series and countries: The heatmaps and scatter plots show that how much data is missing changes over time and also depends on the type of data and the country. Some data types, like “Cereal yield” and “CO2 emissions per person,” are mostly complete, but others have a lot of missing parts. Also, some countries have more complete data than others. Thirdly, the data shows that richer countries generally have higher CO2 emissions per person and use more energy. This matches their more industrial and economic activities. But when you look at CO2 emissions compared to their GDP, richer countries seem to be more efficient. This might mean they are moving towards greener economies or using better technologies.

We also had divergence in income groups and regions. The plots show that climate-related data varies within different income groups and continents. For example, richer countries in the OECD usually do better in certain CO2 measures than non-OECD countries. Also, continents have different patterns, like Europe doing better in cereal yield and CO2 emissions. Furthermore, we had trends even within countries. Looking at specific countries, like Qatar, shows unique situations. Qatar’s high CO2 emissions per person and energy use are due to its oil and gas-based economy and small population. This shows that different countries need different strategies to deal with climate change. In regards to energy source and co2 emissions, perhaps one of the most important in our research, the data suggests that usually, higher energy use per person leads to more CO2 emissions. But there are exceptions, showing that renewable energy can help lower CO2 emissions. This is important for future energy policies and fighting climate change.

In conclusion, our data required careful handling due to its varying completeness and complexity. For example, there are about 233 countries, within each of those 233 countries, there are 18 different Series_name, and within each these Series_name, we had years 1990 to 2011. This was after cleaning the data. Such hierarchical nature of the data required not only cleaning a significant amount of data due to high number of NAs but at the same time made it much more complex to implement graphing techniques like the biplot using PCA. Hence, we had to impute each row of the cleaned dataset. As a result, we were able to reveal significant insights into the relationship between economic development, energy use, and environmental impact, underscoring the diverse challenges faced by different countries in addressing climate change. The findings suggest that targeted, country-specific strategies that consider economic, technological, and environmental factors are essential for effective climate change mitigation and sustainable development.