Relationship Analysis and Dimensional Reduction on Economics

• 

  Circle of Correlation 1

• 

  PCA

• 

  Circle of Correlation 2

• 

  Relationship Analysis Output

• 

  PCA Projection

Inspiration

We were inspired by prior coursework in relationship analysis, as well as prior technical expertise that allowed us to use something as advanced as PCA to bring even more insights than a flat analysis. The real interesting part, or the "crux" of the analysis so to speak, was seeing what contributed to making a company impoverished. The story the different environmental metrics told, the challenges faced by the recorded countries - these were the most captivating part of our analysis.

What it does

We used a relationship analysis to identify the top 20 environmental indicators correlated to GDP Per Capita, then examined deeper with principle component analysis.

How we built it

Python for data manipulation and processing, NumPy and Pandas for numerical operations and dataset handling, Matplotlib and Seaborn for data visualization, Scikit-learn for applying PCA, t-SNE, and UMAP.

Challenges we ran into

Finding an initial dataset that we could link to GDP data and produce worthwhile analysis was challenging. We were able to solve this by finding and cleaning a CSV with embedded GDP and environment data that would serve our purpose

Accomplishments that we're proud of

We would say all of it. We put our best effort in from day one and created a work that we are both proud of. We hope you agree!

What we learned

This exercise was a great way to hone our skills. We both took routes that most interested us, with one handling relationship analysis and the other handling PCA. Not to mention, the project itself gave us a great understanding of how environmental aspects and economic progress can be linked.

What's next for Relationship Analysis and Dimensional Reduction on Economics?

Further analysis, of course! Most likely, a qualitative analysis would be the best option to pursue further studies. We would be fascinated to learn about the causations behind these correlations.

Built With

• matplotlib
• matplotlib-and-seaborn-for-data-visualization
• numpy
• numpy-and-pandas-for-numerical-operations-and-dataset-handling
• pandas
• python
• scikit-learn
• scikit-learn-for-applying-pca
• seaborn
• t-sne