Assignment 6

Data Science: Human Development Index for countries

Introduction

This study will identify the factors which are pivotal in driving Human Development Index for a country. Human Development Index (HDI) is visualised as a function of Economic, Health and Education conditions in a country. Objective of this study is to understand which factors significantly affect HDI and also a view into how HDI is calculated for countries. I will use a model to understand the effects of economic, health and education indicators in HDI calculation. This is pivotal for us to also understand why India is low ranking in HDI. Preprocessing and analysis is done in R (and MS excel just for adjusting data)

Data : Collection and Cleaning

Data was collected from UN data rep. for countries across the world. Factors of interest are estimated using some metrics.Metrics for factors listed above are as follows- Data was cleaned in Excel spreadsheet before importing via simple transpose and deletion of empty columns. Data is imported in R workbook using .read_csv function- Data Summary statistics-

Model

Here, lifeexp,edu and inc represent average life expectancy (in years), mean years of schooling, gross national income (in thousand dollars). Error term is added to account for effects which are not measured by 3 regressors chosen for the analysis. Regression HDI on above mentioned regressors to obtain the OLS estimates of parameters in model- Regression results are as follows- As per the t-statistic calculated from above regression all parameters- intercept and coefficients of metrics are statistically significant.

Results

According to the coefficients estimated by OLS method, a unit increase in life expectancy leads to increase of 0.008 units, a unit increasein mean years of schooling leads to increase of 0.023 units, a $1000 increase in gross national income per capita per year leads to increase of 0.0014 units in HDI index. We can see that HDI is greatly affected by education and least affected by increase in gross national income.Hence, according to Human Development Index metric set by UN education is given greater weight than any other variable Fit of model with individual regressors can be understood from plots below- It can be observed that Education and Health have a good fit but Income is more like a square root function of HDI instead of a staright line.

Further Scope

Model can be modified to a form where the power of income variable is 0.5 instead of 1 to acheive better fit and coefficients.
Analysis used in this assignment is quite lucid and gives a broad picture of how HDI is actually calculated. Further complexities in model can yield much better results and fit.
R-sqaured in current model is 0.9561. Further additions of more variables can improve it- might be dummy variables for various continents or rich/poor countries.