Data Management & Visualization Course

Source code:

STEP 1: univariate graphs.

selected variables:

Life Expectancy

Income per Person

Alcohol Consumption

Life Expectancy:

Center: The distribution is centered at approximately 70-75 years.

Spread: The data is left-skewed, meaning most countries have high life expectancy, but there is a long tail towards the left representing countries with lower health outcomes (ranging down to 45 years).

Income per Person:

Center: The center of the data is located at the very low end of the scale (below $5,000).

Spread: This variable is highly right-skewed. There is a massive concentration of countries at low income levels and a very long "tail" that extends to the right, showing a few outlier countries with extremely high GDP per capita.

Alcohol Consumption:

Center: The distribution is centered around 5-6 liters per capita.

Spread: The data is relatively spread out across the entire range (0 to 15+ liters), but it shows a higher frequency of countries with low-to-moderate consumption levels.

STEP 2: bivariate graph

For the bivariate analysis, Income per Person has been selected as the explanatory variable to investigate its relationship with Life Expectancy, which serves as the response variable.

The goal of this visualization is to determine if a country's economic wealth (represented by GDP per capita) acts as a significant predictor of its population's longevity. By plotting these two variables together, we can observe whether higher income levels correlate with improved health outcomes and identify the strength and direction of this association.

To examine the association between the variables, It has been defined them as follows:

Explanatory Variable (X-axis): incomeperperson. This is the independent variable used to explain the variation in health outcomes. Hypothesis: the country's economic capacity influences its population's longevity.

Response Variable (Y-axis): lifeexpectancy. This is the dependent variable. It represents the outcome or "response" that changes based on the level of income per person.

STEP 1: Run program.

Source Code:

Step 2: Run frecuency / Output.

Rubric items:

1. Program output:

Compared to Assignment 2, the presentation of my frequency tables has changed significantly. In the previous stage, the output displayed raw continuous data, which was difficult to interpret due to the high number of unique values.

For this assignment, I implemented binning (grouping) to collapse these values into four meaningful categories: Low, Medium-Low, Medium-High, and High. This transformation (using quartiles) allows for a much clearer analysis of how the 171 countries are distributed across the economic and health spectrums.

2. The output displays three variables that have been actively managed: Life Expectancy, Income, and Alcohol Consumption. Each has been recoded into categorical frequency tables (Low to High) to provide a clear, standardized distribution of the 171 countries.

3. Summary:

In this stage of the analysis, I managed the data for 171 countries by grouping continuous variables into four categorical quartiles.

Values and Frequencies: For all three variables (Life Expectancy, Income, and Alcohol Consumption), the data is distributed almost equally across four levels: Low, Medium-Low, Medium-High, and High. Each category accounts for approximately 25.14% of the sample. This means that each 'level' contains about 42 to 43 countries, providing a balanced framework for comparison.

Data Management Decisions: I chose to use quartiles (pd.qcut) to collapse the raw, uninterpretable lists of numbers into standardized groups. This allows us to see the relative standing of each nation within the global distribution.

To meet the requirements for this second assignment, we must incorporate a third variable from the Gapminder dataset: alcconsumption.

STEP 1: Run your first program.

Source Code:

Step 2: Run frecuency / Output.

Rubric items:

Program output: The output is organized into clear frequency tables with descriptive headers. Each table includes the calculated frequency and percentage for each interval, including a 'TOTAL' row to confirm the sample size.

Each of the three selected variables is displayed in its respective frequency table: Life Expectancy, Income per person and alcohol consumption.

Each table displays the selected intervals (values), including columns for frequency and percentage to show how often these values occur. Furthermore, row selection was performed to filter out countries with missing data (Step 3 in the code), and a final 'TOTAL' row is included to verify the complete sample size.

Data Interpretation: In this preliminary analysis, Life Expectancy shows a balanced distribution across all intervals because quartiles were used to divide the data (each group representing approximately 25% of the sample). In contrast, the majority of countries fall into the lower levels for both income per person and alcohol consumption.

STEP 1: Data Set Selection I have chosen to work with the Gapminder dataset to explore global trends in health and economics.

STEP 2: Identify a Specific Topic of Interest After looking through the codebook for the Gapminder dataset, I decided that I am particularly interested in public health, specifically life expectancy.

STEP 3: Prepare a Personal Codebook (First Topic) Since Gapminder provides a clear, single indicator for this topic, I added the lifeexpectancy variable to my personal codebook:

Variable Name: lifeexpectancy

Description: 2011 life expectancy at birth (years). The average number of years a newborn child would live if current mortality patterns were to stay the same.

STEP 4: Identify a Second Topic While life expectancy is a good starting point, I needed to determine what factors might influence how long people live across different countries. It strikes me that nations with higher economic resources generally seem to have better healthcare systems and living conditions. Therefore, I decided that I am most interested in exploring the association between a country's wealth and its life expectancy.

STEP 5: Add Second Topic to Personal Codebook I added to my personal codebook the variable reflecting economic development (incomeperperson):

Variable Name: incomeperperson

Description: 2010 Gross Domestic Product per capita in constant 2000 US$. The inflation but not the differences in the cost of living between countries has been taken into account.

STEP 6: Literature Review To understand the existing research on this topic, I conducted a brief literature review:

Search Terms Used: I used Google Scholar with the exact search terms: "GDP per capita and life expectancy", "economic development and mortality rates", and "income inequality and public health".

Summary of Findings: The literature consistently shows a strong, positive relationship between national income and life expectancy, a pattern famously known as the "Preston Curve". Researchers highlight that higher income per capita allows countries to invest more heavily in healthcare infrastructure, sanitation, and nutrition. Interestingly, studies also show that this relationship is non-linear: gains in life expectancy are massive when moving from extreme poverty to middle-income, but the curve flattens out among the wealthiest nations, where lifestyle factors play a larger role.

References Used:

Preston, S. H. (1975). The changing relation between mortality and level of economic development. Population Studies, 29(2), 231-248.

Deaton, A. (2003). Health, inequality, and economic development. Journal of Economic Literature, 41(1), 113-158.

STEP 7: Research Question & Hypothesis

Research Question: Is there a direct and positive association between a country's income per person (incomeperperson) and the life expectancy of its inhabitants at birth (lifeexpectancy)?