Data Science Project

This choropleth map shows the relationship between social mobility and income inequality. The social mobility data was scraped using Python from Wikipedia (derived from World Economic Forum). The data will refresh automatically when the code is run as I used the World Bank Data python library to download data on the Gini coefficient from the World Bank API. There were many gaps in the data for some countries and duplicates for others, so I used data from 2010 only and used the most recent values in Python. I merged the datasets in pandas, converted the TopoJSON file to a GeoJSON, and uploaded the file to Kepler.

The map shows that countries with a high level of inequality generally have a low social mobility index score. There seems to be a large disparity between continents; for instance, Europe typically has lower levels of inequality and higher levels of social mobility, with South America being the opposite. Interestingly, there are some anomalies with countries such as Bangladesh and Tunisia displaying high levels of inequality and social mobility.

Social mobility is most often quantitatively measured in terms of changes in economic mobility, such as changes in income or wealth. In the United States, the share of total net worth held by the top 1% has gradually increased over the last 30 years. In contrast, the total net worth growth held by the bottom 50% has been significantly more volatile.

I used a loop over an API to obtain the data from Federal Reserve Economic Data (FRED), making it replicable.

The data for these charts came from the OECD. The OECD has a poor API integration making it difficult to get the data using XML requests, so instead, I downloaded the data locally. The OECD measures educational performance using PISA scores. PISA is an international survey providing a cross-national comparison of student performance. The PISA assessment also provides an index of students' economic, social, and cultural status (ESCS) by combining parents' educational level, parents' labour market status, and household possessions. The higher the ESCS score, the more advantaged that student is.

Across all OECD countries, students with high ESCS scores perform significantly better than students with low scores. The difference is especially large in countries such as Israel and Luxembourg, where the difference between PISA scores between the top and bottom quarters of ESCS is greater than 120 points.

This Trellis chart shows that across OECD countries, adults are more likely to have high levels of educational attainment when their parents also have attained high levels of education. On average across OECD countries, 68% of adults who have at least one parent with tertiary education have also attained tertiary education, compared to 39% among those with parents with an upper secondary education, and 22% among those with parents with at most below upper secondary education.

Housing is thought to either block or expands people’s access to opportunities for upward mobility. There may be a weak correlation between social mobility and home ownership rate in Europe, where countries with lower home ownership levels exhibit higher social mobility levels.

This choropleth map was produced by utilising BeautifulSoup to scrape two datasets from Wikipedia and merge them through pandas, making it fully easily updatable. I used Wikipedia due to the lack of sources that provide data for all the countries required. I tried using Folium to make the Kepler maps visible on a mobile phone, but since it lost its interactivity, it was not used.

I regressed social mobility with GDP/Capita and Life Expectancy. I used the World Bank Data python library to download data, enabling new data to be downloaded annually upon release. I had to reindex the tables and rename columns to merge the data smoothly before cleaning commas, currencies, and quotation marks.

The R-Squared values of the regressions is 0.37 suggesting there is a correlation, albeit relatively weak, between social mobility and both GDP/Capita and Life Expectancy.