We humans have been making maps of the world since the beginnings of our culture, from the babylonian Imago Mundi to the latest big data based digital maps.
Here a new world map is created from a geometric interpretation of data:
A high-dimensional space is constructed from socio-economic data about countries using different data sources.
Each country is a point in space and each metric is a dimension (around 100 dimensions for the current map).
Using dimensionality reduction (TSNE), clustering (DBSCAN) and topological data analysis (Kepler mapper), the points in high-dimensional space are projected and embedded in two dimensions.
The result is a flat map that approximates the shape of data in high-dimensional space.
A new world map with different borders and regions emerges. An image of a different geography, not the physical geography of rivers, mountains and deserts, but a geography of people, lives and human activity.
Created by Christian Parsons
Many thanks to the people that created the open-source tools and data sources used in this project
Tools:
kepler-mapper.scikit-tda.org scikit-learn.org pandas.pydata.org d3js.org lodash.comData sources:
World Data Visualization Prize (dataset: What Makes A "Good" Government?) Social Progress Index Economic Complexity Index