Exploring Poverty with Principal Component Analysis

Project Summary

This Data Expedition seeks to introduce students to statistical analysis in the field of international development. Students construct a index of wealth/poverty based on asset holdings using four datasets collected under the umbrella of the Living Standards Measurement Survey project at the World Bank. We selected countries to represent different continents with comparable and recent survey data: Bulgaria (2007), Tajikistan (2009), Tanzania (2010-2011), and Panama (2008).

First, we construct an index of wealth based on household assets in the different countries using Principle Components Analysis. Once a poverty index is constructed, students seek to understand what the main drivers of wealth/poverty are in different countries. We include variables for health, education, age, relationship to the household head, and sex. Students then use regression analysis to identify the main drivers of poverty in different countries.

Themes and Categories
Year
2018

Graduate Students: Claire Le Barbenchon (claire.lebarbenchon@duke.edu) and Federico Ferrari (federico. ferrari@duke.edu)

Faculty: Mine Çetinkaya-Rundel

Course: STA 112FS: Data Science

Guiding Questions

This project seeks to help students answer some key questions in international development: How can we get measures of poverty that do not involve income? What are some of the demographic and human capital variables that can help explain wealth levels?

In developing countries, income may not be an ideal measure of well-being. In agricultural societies, income is tied to harvest, and thus is sensitive to the season in which the survey is administered. Furthermore, income often does not capture in-kind revenue, and is a particularly poor measure for households that practice subsistence agriculture. For this reason, development experts will often use indices of poverty based on household assets, to better understand well-being levels. Household asset ownership does not vary seasonally, and is not tied to payment-type, making it a more stable and often preferred measure of poverty.

Conceptual and Statistical Questions:

  • Why we may want to create a poverty or wealth index?
  • How to use Principal Component Analysis (PCA) to construct indices
  • How to interpret PCA results and compare across countries
  • Using regression to understand the factors that affect wealth across countries
  • Using and interpreting interaction terms

The plots below show the distribution of wealth in each of the four countries. They can be used to answer descriptive questions about inequality.

Figure 1: Distribution of first component
Figure 1: Distribution of first component
Figure 2: Boxplot of first component
Figure 2: Boxplot of first component

Data

All the data sets used in this project were collected under the umbrella of the Living Standards Measurement Survey project at the World Bank. This large initiative has collected survey data from developing countries since the 1980s (using comparable survey instruments) to better understand questions of health, education, poverty, employment and other indicators of well-being around the world.

While this data is public use, individuals must register and download the data at http://microdata.worldbank. org/index.php/catalog/central

We created a simplified data set for instructional purposes, randomly selecting 2,500 observations, in order to make sure that sample sizes were uniform across countries.

household_dummy.csv

Dimensions of the dataset are 2,500 observations per country, for 4 countries, and 20 variables including: an ID variable for the individual and the household; and a country variable; the following demographic variables: age, gender, marital status, relationship to household head; education; a health proxy variable (hospitalization in past 12 months); water access; and ten household assets (refrigerator, tv, bike, motorbike, computer, car, video, stereo, stove, sewing machine). Thus the full dimensions are 10,000x20, including the variable for the country.

Bibliography

The World Bank, Living Standards Measurement Study LSMS (2007). Bulgaria Multitopic Household Survey 2007 [BGR_2007_MTHS_v01_M]. Retrieved from http://microdata.worldbank.org/index.php/catalog/ 2273/study-description

The World Bank, Living Standards Measurement Study - Integrated Surveys on Agriculture (2010-2011). Tanzania - National Panel Survey 2010-2011, Wave 2 [TZA_2010_NPS-R2_v01_M]. Retrieved from http: //microdata.worldbank.org/index.php/catalog/1050

The World Bank, Living Standards Measurement Study LSMS (2008). Panama - Encuesta de Niveles de Vida 2008 [PAN_2008_ENV_v01_M]. Retrieved from http://microdata.worldbank.org/index.php/catalog/70

Tajikistan Statistical Agency, Living Standards Measurement Study LSMS (2009). Tajikistan - Living Standards Survey 2009 [TJK_2009_TLSS_v01_M]. Retrieved from http://microdata.worldbank.org/index. php/catalog/73%5Bc1%5D

Related Projects

KC and Patrick led two hands-on data workshops for ENVIRON 335: Drones in Marine Biology, Ecology, and Conservation. These labs were intended to introduce students to examples of how drones are currently being used as a remote sensing tool to monitor marine megafauna and their environments, and how machine learning can be used to efficiently analyze remote sensing datasets. The first lab specifically focused on how drones are being used to collect aerial images of whales to measure changes in body condition to help monitor populations. Students were introduced to the methods for making accurate measurements and then received an opportunity to measure whales themselves. The second lab then introduced analysis methods using computer vision and deep neural networks to detect, count, and measure objects of interest in remote sensing data. This work provided students in the environmental sciences an introduction to new techniques in machine learning and remote sensing that can be powerful multipliers of effort when analyzing large environmental datasets.

This two-week teaching module in an introductory-level undergraduate course invites students to explore the power of Twitter in shaping public discourse. The project supplements the close-reading methods that are central to the humanities with large-scale social media analysis. This exercise challenges students to consider how applying visualization techniques to a dataset too vast for manual apprehension might enable them to identify for granular inspection smaller subsets of data and individual tweets—as well as to determine what factors do not lend themselves to close-reading at all. Employing an original dataset of almost one million tweets focused on the contested 2018 Florida midterm elections, students develop skills in using visualization software, generating research questions, and creating novel visualizations to answer those questions. They then evaluate and compare the affordances of large-scale data analytics with investigation of individual tweets, and draw on their findings to debate the role of social media in shaping public conversations surrounding major national events. This project was developed as a collaboration among the English Department (Emma Davenport and Astrid Giugni), Math Department (Hubert Bray), Duke University Library (Eric Monson), and Trinity Technology Services (Brian Norberg).

Understanding how to generate, analyze, and work with datasets in the humanities is often a difficult task without learning how to code or program. In humanities centered courses, we often privilege close reading or qualitative analysis over other methods of knowing, but by learning some new quantitative techniques we better prepare the students to tackle new forms of reading. This class will work with the data from the HathiTrust to develop ideas for thinking about how large groups and different discourse communities thought of queens of antiquity like Cleopatra and Dido.

Please refer to https://sites.duke.edu/queensofantiquity/ for more information.