Sartirano Daniele, Kalimeri Kyriaki, Cattuto Ciro, Delamónica Enrique, Garcia-Herranz Manuel, Mockler Anthony, Paolotti Daniela, Schifanella Rossano
ISI Foundation, Turin, Italy.
UNICEF, New York, NY, United States.
Front Big Data. 2023 Feb 21;6:1054156. doi: 10.3389/fdata.2023.1054156. eCollection 2023.
Accurate relative wealth estimates in Low and Middle-Income Countries (LMICS) are crucial to help policymakers address socio-demographic inequalities under the guidance of the Sustainable Development Goals set by the United Nations. Survey-based approaches have traditionally been employed to collect highly granular data about income, consumption, or household material goods to create index-based poverty estimates. However, these methods are only capture persons in households (i.e., in the household sample framework) and they do not include migrant populations or unhoused citizens. Novel approaches combining frontier data, computer vision, and machine learning have been proposed to complement these existing approaches. However, the strengths and limitations of these big-data-derived indices have yet to be sufficiently studied. In this paper, we focus on the case of Indonesia and examine one frontier-data derived Relative Wealth Index (RWI), created by the Facebook Data for Good initiative, that utilizes connectivity data from the Facebook Platform and satellite imagery data to produce a high-resolution estimate of relative wealth for 135 countries. We examine it concerning asset-based relative wealth indices estimated from existing high-quality national-level traditional survey instruments, the USAID-developed Demographic Health Survey (DHS), and the Indonesian National Socio-economic survey (SUSENAS). In this work, we aim to understand how the frontier-data derived index can be used to inform anti-poverty programs in Indonesia and the Asia Pacific region. First, we unveil key features that affect the comparison between the traditional and non-traditional sources, such as the publishing time and authority and the granularity of the spatial aggregation of the data. Second, to provide operational input, we hypothesize how a re-distribution of resources based on the RWI map would impact a current social program, the Social Protection Card (KPS) of Indonesia and assess impact. In this hypothetical scenario, we estimate the percentage of Indonesians eligible for the program, which would have been incorrectly excluded from a social protection payment had the RWI been used in place of the survey-based wealth index. The exclusion error in that case would be 32.82%. Within the context of the KPS program targeting, we noted significant differences between the RWI map's predictions and the SUSENAS ground truth index estimates.
在低收入和中等收入国家（LMICs）中，准确的相对财富估计对于帮助政策制定者在联合国设定的可持续发展目标指导下解决社会人口不平等问题至关重要。传统上，基于调查的方法被用于收集有关收入、消费或家庭物质商品的高度细化数据，以创建基于指数的贫困估计。然而，这些方法仅涵盖家庭中的人员（即在家庭样本框架内），不包括流动人口或无家可归的公民。已经提出了结合前沿数据、计算机视觉和机器学习的新方法来补充这些现有方法。然而，这些大数据衍生指数的优势和局限性尚未得到充分研究。在本文中，我们聚焦印度尼西亚的案例，研究一个由Facebook数据为善倡议创建的前沿数据衍生相对财富指数（RWI），该指数利用Facebook平台的连接数据和卫星图像数据，为135个国家生成相对财富的高分辨率估计。我们将其与根据现有的高质量国家级传统调查工具（美国国际开发署开发的人口与健康调查（DHS）以及印度尼西亚全国社会经济调查（SUSENAS））估计的基于资产的相对财富指数进行比较。在这项工作中，我们旨在了解前沿数据衍生指数如何用于为印度尼西亚和亚太地区的扶贫项目提供信息。首先，我们揭示影响传统和非传统数据源之间比较的关键特征，例如发布时间和权威性以及数据空间聚合的粒度。其次，为了提供操作建议，我们假设基于RWI地图重新分配资源将如何影响当前的社会项目——印度尼西亚的社会保护卡（KPS），并评估其影响。在这个假设情景中，我们估计了符合该项目条件的印度尼西亚人的百分比，如果使用RWI代替基于调查的财富指数，这些人将被错误地排除在社会保护支付之外。在这种情况下，排除误差将为32.82%。在KPS项目目标设定的背景下，我们注意到RWI地图的预测与SUSENAS地面真值指数估计之间存在显著差异。
