World Population

Predicting Yearly Change

Project
Overview

In this project, we are attempting to use 2020 world population data to see what, if any, factors can be used to predict the percentage change in a given country's population from 2019 to 2020.

CHECK IT OUT

Visualizations

Some positive and some negative correlations were observed between the features of this dataset, as seen in the lighter-most and darker-most squares pairing two separate features.

Median Age vs Yearly Population Change

The data show a negative correlation between the median age of a country and the rate of population growth. K-means clustering was used in this scatter plot to try to find additional groups in the data. Each cluster is shown as a different color and it is clear that clustering did not help show any unknown groups.

Fertility Rate vs Yearly Population Change

Here the data show a positive correlation between a higher fertility rate and the growth of a countrys population. K-means clustering was used here as well, again with no useful results.

Correlating both Fertility Rate and Median Age with Yearly Population Change

A country with a higher fertility rate is likely to have a lower median age, and thus again correlate with a higher percentage change in population.

Tableau Visualizations

Yearly Change Heatmap

Machine Learning Outcomes

Linear Regression

Neural Network

Random Forest Regressor

Success

Analysis

Other than median age and fertility rate, there were not any features in the dataset that had a notable relationship with the percentage of growth or decline in a given country from 2019 to 2020. We did find a negative correlation between a higher median age and a smaller percentage change in population growth, as well as a positive correlation between a higher fertility rate and a larger percentage change in growth. These two findings make sense, as it seems likely that a country with a higher fertility rate would have more young citizens and thus a lower median age. These data are perhaps not surprising.

There is, however, a caveat. 2020 was the year that the COVID-19 pandemic erupted worldwide. This affected many aspects of people’s lives across the globe, including plenty of the features shown in our dataset. This makes 2020 an outlier year in many factors and, on its own, may appear a strange period to analyze. However, used in relation to years where population data conforms more to the mean, this single year could be an interesting point of comparison.

Group