African GDP Outliers

Initializing live version
Download to Desktop

Requires a Wolfram Notebook System

Interact on desktop, mobile and cloud with the free Wolfram Player or other Wolfram Language products.

For the most part, African national GDP data for individual countries for 1970–2008 correlates very strongly, with pairs of data showing a correlation of more than 0.9. Connecting such pairs of "strongly" correlated countries together creates a graph containing most countries. The indirectly correlated countries are those that are connected to each other only via other vertices. Countries that correlate "weakly" are isolated vertices. It may be interesting to examine the GDP data with respect to historic events like a change of government, civil wars, and so on.


Also shown are correlations for land areas and population.


Contributed by: Trevor Cole (August 2012)
Open content licensed under CC BY-NC-SA



Most of the GDPs of African countries from 1970–2008 are highly correlated. Using the built-in Mathematica function CountryData, the GDPs are correlated. You can choose the countries with the largest population, the largest land area, or the largest GDP. You can also choose how many countries to view. The GDP data that does not correlate strongly with other countries (correlation < 0.9) is identified using a copy of the graph theory technique described in [1].

The GDPs of pairs of countries can be compared by clicking the "compare GDP" button and selecting two countries.

The "graph" button shows how the GDP data of a selected country correlates with data from other countries. The graph (see below) is constructed linking pairs of countries whose GDPs showed a greater than 0.9 correlation.

The correlation of a country with itself is always equal to 1 and is denoted as an "absolute" correlation.

If the correlation of the second country's data with the first country's data is greater than 0.9, it is deemed to be "strongly" correlated and is directly connected to the first country in the graph.

An "indirectly" correlated country is one whose correlation with the selected country is less than 0.9 but is linked to the selected country via other countries.

A "weakly" correlated country is one whose correlation with the selected country is less than 0.9 and is not linked to the selected country via other countries.

The "distance" between two countries is the number of graph vertexes traversed to get from one country vertex to the other. If there is no connection between countries, then the distance is infinity; the distance from a country to itself is zero.


[1] S. Chen, "Graph Theory and Finance in Mathematica," Wolfram Blog (blog). (Jun 1, 2012)

Feedback (field required)
Email (field required) Name
Occupation Organization
Note: Your message & contact information may be shared with the author of any specific Demonstration for which you give feedback.