The Problem

Economic Inequality, also known as the wealth or income gap, is a growing concern that's associated with poverty, inequality of opportunity, and many other issues ranging from social to environmental. Some research has shown that in societies of high economic inequality, an "envy up, scorn down" mentality bias exists, meaning that members look more negatively upon those of lower economic class and more positively upon those of higher economic class than would be the case in a society of lesser economic inequality. I wanted to see if this was true.

The Solution

To answer the question, I mined Twitter data to see how sentiment towards the rich and poor correlated with regions of high and low economic inequality. First, using the Twitter API, I collected over 2000 Tweets from regions in the U.S. of the 10 highest and 10 lowest economic inequality. This was measured by an index similar to the Gini Coefficient. The Tweets either contained words relating to high or low economic class like "rich," "affluent," "wealthy" or "poor," "broke," "cheap," etc. The Tweets were then analyzed with the IBM Watson text analysis API and given a sentiment score. The level of inequality by sentiment was measured as the range between the averages among positive sentiment for the rich and negative sentiment for the poor. Finally, a correlation test was performed against this measurement and the official measurement. In the end after various tests, results upheld the origin hypothesis.

See the PowerPoint presentation for details.