Topic: Poverty Assessment and Analysis

Article 1: https://royalsocietypublishing.org/doi/10.1098/rsif.2017.0401 Authors: Jeremiah J. Nieves, Forrest R. Stevens, Andrea E. Gaughan, Catherine Linard, Alessandro Sorichetta, Graeme Hornby, Nirav N. Patel and Andrew J. Tatem

Nieves Jeremiah J., Stevens Forrest R., Gaughan Andrea E., Linard Catherine, Sorichetta Alessandro, Hornby Graeme, Patel Nirav N. and Tatem Andrew J. 2017 “Examining the correlates and drivers of human population distributions across low- and middle-income countries” J. R. Soc. Interface.14

This article describes the drivers and correlates of human population distribution across low/middle income countries. Specifically, the article hypothesizes that geographic factors including climate, topography, and transportation links among others will have a significant correlation with human population distribution which in turn can shed light on poverty assessment and analysis as the UN predicted that human population growth would occur in significantly more impoverished areas across the world. The authors proceed to use different data analysis techniques using machine learning algorithms to quantify the importance of these factors across 32 low and middle income countries. Poverty assessment and analysis becomes extremely complicated as it can involve every level of society ranging from internal factors such as local political and economic policy to global factors such as geography and international trade. Thus, using data science to analyze the influence of geographical factors in driving poverty through human population distributions becomes an attractive solution to understanding the fundamentals surrounding poverty analysis. After conducting each analysis, the authors concluded that urban areas and climatic/environmental factors were the most significant in predicting population density and thus allows for quantifiable analysis of poverty within regions. Their analysis sheds light on the identification of poverty by showing how variables like urbanization extent, elevation, resource proximity, crop/vegetation land cover, and more can drive population growth. Ultimately, if scientists can better understand these factors that contribute to population growth, high-resolution mapping of poverty and construction/preparation for future development scenarios can be achieved.

This article relates to Sen’s definition of freedom by describing a solution for a removal of a major unfreedom which is poverty. In this instance, an analysis and potential solution is depicted to better understand and predict poverty through human population distributions using data science techniques. Sen says that the removal of major sources of unfreedoms is crucial for development and understanding how various geographical factors contribute to human distribution is conducive to that goal of mitigating an unfreedom and promoting development.

Targeting poverty through quantifiable and analytical means is being addressed by the authors research. This is an important dimension of human development as 9.2% of the world is in extreme poverty while a significantly larger portion lives in need of at least some basic necessities. By better understanding factors that contribute to human population growth, it becomes easier to identify development implementation for the future.

No poverty, Industry/Innovation/Infrastructure, and Sustainable Cities and Communities are all sustainable development goals that are related to the article. If human population growth can be better understood, development methods for reducing poverty, effective infrastructure usage, and maintaining living standards in cities and communities can be achieved.

The authors sample census data from 32 low and middle income countries in 4 regions around the world at an average spatial resolution of 100 km^2 or below. In their analysis, they used intensity of night-time lights, energy productivity of plants, topographic elevation and slope, climatic factors, type of land cover and presence/absence of roads, water features, human settlements and urban areas, protected areas, and locations of points of interest (POIs) and facilities such as health centres and schools as predictive covariates of population growth.

The authors employ random forest-based population models to conduct their analysis. They describe RF’s as a non-parametric, nonlinear statistical method that falls within a category of machine-learning methods known as ‘ensemble methods’. Ensemble methods take individual decision trees that are considered ‘weak learners’ and combine them to create a ‘strong learner’.

The authors are seeking to find a relationship between geographic factors and human population densities. Human population densities are an important pattern for scientists to understand as they provide insight concerning human development through predicting and alleviating poverty.

The scientists are seeking to answer the question, “What factors correlate and drive human population distributions in low and middle income countries?”

Article 2: https://royalsocietypublishing.org/doi/full/10.1098/rsif.2016.0690

Steele Jessica E., Sundsøy Pål Roe, Pezzulo Carla, Alegana Victor A., Bird Tomas J., Blumenstock Joshua, Bjelland Johannes, Engø-Monsen Kenth, de Montjoye Yves-Alexandre, Iqbal Asif M., Hadiuzzaman Khandakar N., Lu Xin, Wetter Erik, Tatem Andrew J. and Bengtsson Linus 2017 “Mapping poverty using mobile phone and satellite data” J. R. Soc. Interface.142016069020160690

This article seeks to find a modern and more effective method of mapping poverty by taking advantage of phone and satellite data. In this study, the authors base their claim on the idea that current metrics of poverty are predicated on outdated census data, and instead employ data science techniques using commonly available public and private data from mobile phones and satellites in hopes of providing insight into the spatial distribution of poverty. By using new techniques to update the accuracy of poverty maps, it becomes possible to to understand the causes of poverty as well as methods of poverty alleviation for development purposes. Specifically, the authors use remote sensing and geographic information system (RS) data as well as call detail records (CDR) to accurately map poverty. Poverty mapping is an integral part of poverty assessment and analysis as it allows scientists to accurately identify what exact areas or regions will be affected by lack of resources while also providing a streamlined framework for allievations concerning development purposes. Conclusively, the researchers successfully used RS and CDR data to compile poverty maps with significant accuracy as their models were able to successfully produce high-resolution maps that were based on various data factors.

Based on Sen’s definition of development as freedoms and the mitigation of unfreedoms, this article relates to the idea of alleviating poverty as an unfreedom through increased awareness and knowledge of poverty as a result of compiling poverty maps using RS and CDR data. By having access to modern data that can be updated in real time as opposed to census data, poverty mapping from RS and CDR data will prove to be a valuable tool that can be used for efficient development purposes and alleviation.

Mapping poverty in hopes of providing effective methods for poverty reduction is a development dimension addressed in this article. If poverty is mapped effectively, there is a huge upside potential for scientists and development experts to provide methods of alleviation and resource management in an extremely efficient manner. In addition, using data that can be accessed and updated at any time proves to be vital as it allows for the process to be streamlined and accurate at every level.

No poverty is a sustainable development goal addressed by the article. The authors of the article cite this SDG and hope that their research can support the eradication of poverty. In order to do this, information is necessary to determine where poverty occurs and this can be made possible through the compilation of poverty maps using the RS and CDR data.

The authors employ three primary sources of data as well as other secondary sources of data. First, they used three geographically referenced datasets representing asset, consumption and income-based measures of wellbeing in Bangladesh. Then, CDR, or call detail records, were generated from mobile phone metadata spanning four months using Bangladesh’s largest mobile phone network that covers 90% of the land and 99% of the population. CDR features range from metrics such as basic phone usage, top-up patterns, and social network to metrics of user mobility and handset usage. In addition, 25 raster and vector datasets were processed into remote sensing/geographic (RS) covariates at a spatial resolution of 1 km^2. These data acted as environmental and physical metrics that were likely to be associated with human welfare and included vegetation indices, night-time lights, climatic conditions, and distance to roads or major urban areas.

The authors first had to identify a set of predictors that were most suitable for modelling the 3 datasets for poverty. To do this, they employed non-spatial generalized linear models that were trained on 80% of randomly selected data to prevent overfitting. Then, to generate the actual poverty mapping, they used the models from the previous step through hierarchical Bayesian geostatistical models (BGMs) to predict the poverty metrics. The authors chose BGMs as they offer several advantages including straightforwardly imputing missing data and estimating uncertainty in the predictions as a distribution around each estimate.

The authors are investigating how poverty is distributed using mobile phone data. The human development dimension of poverty and its implications specifically are crucial to understand if poverty alleviation is to be achieved.

The scientific question the authors are trying to answer is “Can mobile phone and satellite data be used to accurately map poverty?”

Article 3: https://openknowledge.worldbank.org/bitstream/handle/10986/31267/WPS8735.pdf?sequence=1&isAllowed=y

Authors: Utz Pape, Philip Wollburg

Pape, Utz, and Philip Wollburg. “Estimation Of Poverty In Somalia Using Innovative Methodologies”. 2019. World Bank, Washington, DC, doi:10.1596/1813-9450-8735

Readily available data in Somalia is extremely deprived, thus making any statistical calculations or estimates of poverty very difficult. As a result, the World Bank implemented the second wave of the Somali High Frequency Survey in hopes of estimating national poverty indicators. This article explains how the survey overcomes the lack of statistical infrastructure and specific insecurity in Somalia through methodological and technological adaptations in four areas. Primarily, geospatial techniques and high-resolution imagery were used to model the spatial population distribution, build a probability-based population sampling frame, and generate enumeration areas to overcome the lack of a recent population census. Second, the survey adapted logistical arrangements, sampling strategy using micro-listing, and questionnaire design to limit time on the ground based on the Rapid Consumption Methodology. Third, correlates were derived from satellite imagery and other geospatial data to estimate poverty. Fourth, special sampling techniques were employed to capture the non stationary nature of the nomadic population. In an area like Somalia where infrastructure and means to collect data is lacking, it becomes ever harder to measure indicators that shed light on the status of poverty in hopes of development purposes. Assessing and analyzing poverty is crucial to implement development and that process is rooted in understanding/collecting the data. Scientists must find innovative ways and create alternatives of collecting data when faced with challenges like the lack of indicators. Thus, the survey proposes to overcome that challenge in the variety of ways described above and understanding how the survey does so is vital for development.

Sen’s definition of development involves roles of freedom as well as evaluative systems for development. By using the survey to compile and gain a better understanding of data where there is a lack of statistical infrastructure, it becomes possible for scientists and development experts to effectively evaluate poverty and means of poverty reduction.

Poverty and evaluative measures (economic, geographic etc.) are dimensions that are measured in the article. The survey hopes to glean an accurate measure of these various factors in hopes of accurate analysis of poverty for development implementation .

No poverty, Good Health/Wellbeing, and Industry/Innovation/Infrastructure are being addressed in this article as the aforementioned survey hopes to identify these factors in an otherwise data lacking landscape for the measurement of poverty.

The first Somali High Frequency Survey was used as a baseline data set for monitoring poverty and other key statistical indicators. In addition, GIS data sets like Pre‐war region boundaries; (ii) IDP settlement boundaries; (iii) Urban area boundaries; (iv) Rural settlement boundaries; (v) Security assessment access map were used to demarcate strata boundaries. Population estimates were derived from the 2015 World Pop data sets. Furthermore, data sets were created using a number of existing indicators like ‘Visible Infrared Imaging Radiometer Suite’ (VIIRS) and Defense Meteorological Satellite Program (DMSP) (satellite data) in hopes of accurately estimating various economic, geographic, population, and poverty numbers.

The authors used a variety of data science techniques to create an accurate data set in an otherwise statistically dry region in hopes of assessing poverty and other economic/geographical factors. They first used geospatial techniques and high‐resolution imagery to model the spatial population distribution, build a probability‐based population sampling frame, and generate enumeration areas in an effort to overcome the lack of a recent population census. Then, logistical arrangements, sampling strategy, and questionnaire design to limit time on the ground were adopted to ensure the safety for those who worked in the field. In addition, a consumption model was assigned to households in areas without a binding security and time constraint, with only the covariates of consumption administered to households in insecure areas to measure consumption and poverty. Ultimately, sampling was conducted using the aforementioned process, data was collected using the limited time on the ground method, the consumption aggregate was compiled, and an overview of poverty in the area was computed.

The authors are investigating how poverty and different economic measures like consumption can be assessed in a data-lacking zone. These dimensions of human development are vital for understanding how poverty can be measured and alleviated.

The scientists are seeking to answer the question “How were poverty and other economic factors measured in a country like Somalia with relatively no existing usable data?”

Article 4: https://openknowledge.worldbank.org/handle/10986/31190

Pape, Utz, and Luca Parisotto. “Estimating Poverty In A Fragile Context: The High Frequency Survey In South Sudan”. 2019. World Bank, Washington, DC, doi:10.1596/1813-9450-8722. Accessed 10 Mar 2021.

Authors: Utz Pape, Luca Parisotto This article attempts to describe the design and analysis of the Sudan High Frequency Survey. Accurately depicting economic factors and poverty in Sudan was extremely difficult as a result of the civil conflict. Thus, the high frequency survey attempts to overcome this challenge in a variety of ways including technological and methodological innovations to establish a reliable system of data collection and obtain valid poverty estimates. Accurately assessing and analyzing poverty in a conflict-ridden region like Sudan is not only difficult but crucial in determining how development can be achieved. Having access to information regarding poverty and economic factors allows scientists to quickly and efficiently allocate resources and other means of poverty alleviation.

This article relates to Sen’s definition of through evaluations and limiting unfreedoms (poverty). The high frequency survey attempts to evaluate various economic, political, and geographical factors in hopes of assessing poverty. This allows scientists to better understand what methods of development will be the most effective in combating poverty.

Poverty and evaluative measures are development dimensions expressed in the article. The survey hopes to use numerous methods of evaluation for these factors in hopes of analyzing and assessing the extent to which poverty afflicts individuals in the country, especially amidst a civil conflict.

No poverty, and Industry/Innovation/Infrastructure are SDGs in relation to the article as the survey attempts to quantify these in an area where compiling data to do so is already difficult.

Existing spatial data-sets and custom‐derived spatial data with geo‐referenced poverty estimates were used to obtain the estimates produced by the survey.

The survey used an expansion of cellular networks in Sudan to build a near-real time monitoring system where data uploaded daily to a dedicated server and checked for consistency. In addition, it used a variety of methods and techniques to produce accurate results. First, poverty was measured using a standardized methodology where a welfare measure was compared to a preexisting poverty line. Consumption was the welfare measure chosen to compare, and thus the researchers had to use different techniques to obtain a consumption number. Then, a poverty line was calculated using a variety of old and new data in tandem to compare to consumption. Ultimately, the survey employed a method of within‐survey imputation, but instead of imputing the totality of consumption in certain areas based on data from other areas it imputed a randomly different fraction of consumption across all enumeration areas covered in the survey.

The authors are investigating how poverty can be estimated in Sudan. This human development pattern of poverty is important in determining how poverty is distributed, especially in a war-torn country like Sudan.

The scientists are hoping to answer the question “How were poverty and other economic factors measured in a post-conflict country like Sudan where there is little access to reliable data?”

I think that South Asia or Africa are ideal regions to study poverty assessment and analysis as data in these regions is hard to obtain and thus development implementation is difficult to quantify. However, these regions are also heavily afflicted by poverty so they are in the most need of poverty alleviation and development.