Combining disparate data sources for improved poverty prediction and mapping

Pokhriyal, Neeti; Jacques, Denis

doi:10.1073/pnas.1700319114

Cited by 151 publications

(93 citation statements)

References 42 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Blumenstock [53] showed how an individual's historical record of mobile phone use can infer his/her socio-economic levels using CDRs. Pokhriyal and Jacques [54] and Steele, et al [55] both explored this topic further by proposing Bayesian frameworks to combine disparate data sources (typically mobile phone and Earth Observation data) for improved poverty prediction and mapping.…”

Section: Urban Population Susceptibilitymentioning

confidence: 99%

Mobile Phone Data for Urban Climate Change Adaptation: Reviewing Applications, Opportunities and Key Challenges

Dujardin

Jacques²,

Steele

et al. 2020

Sustainability

View full text Add to dashboard Cite

Climate change places cities at increasing risk and poses a serious challenge for adaptation. As a response, novel sources of data combined with data-driven logics and advanced spatial modelling techniques have the potential for transformative change in the role of information in urban planning. However, little practical guidance exists on the potential opportunities offered by mobile phone data for enhancing adaptive capacities in urban areas. Building upon a review of spatial studies mobilizing mobile phone data, this paper explores the opportunities offered by such digital information for providing spatially-explicit assessments of urban vulnerability, and shows the ways these can help developing more dynamic strategies and tools for urban planning and disaster risk management. Finally, building upon the limitations of mobile phone data analysis, it discusses the key urban governance challenges that need to be addressed for supporting the emergence of transformative change in current planning frameworks.

show abstract

Section: Urban Population Susceptibilitymentioning

confidence: 99%

Mobile Phone Data for Urban Climate Change Adaptation: Reviewing Applications, Opportunities and Key Challenges

Dujardin

Jacques²,

Steele

et al. 2020

Sustainability

View full text Add to dashboard Cite

show abstract

“…In addition, the length of time that the mobile phone number is used can reflect his/her economic situation, and this is positively related to the economic conditions of the individual [29]. Economic status is an important indicator of P2P loan default prediction [30].…”

Section: Hypothesis Developmentmentioning

confidence: 99%

Credit Scoring Using Machine Learning by Combing Social Network Information: Evidence from Peer-to-Peer Lending

Niu

Ren

2019

Information

View full text Add to dashboard Cite

Financial institutions use credit scoring to evaluate potential loan default risks. However, insufficient credit information limits the peer-to-peer (P2P) lending platform’s capacity to build effective credit scoring. In recent years, many types of data are used for credit scoring to compensate for the lack of credit history data. Whether social network information can be used to strengthen financial institutions’ predictive power has received much attention in the industry and academia. The aim of this study is to test the reliability of social network information in predicting loan default. We extract borrowers’ social network information from mobile phones and then use logistic regression to test the relationship between social network information and loan default. Three machine learning algorithms—random forest, AdaBoost, and LightGBM—were constructed to demonstrate the predictive performance of social network information. The logistic regression results show that there is a statistically significant correlation between social network information and loan default. The machine learning algorithm results show that social network information can improve loan default prediction performance significantly. The experiment results suggest that social network information is valuable for credit scoring.

show abstract

“…One of the first was [13], which combined a Support Vector Machine (SVM) predictor and CDR data for 500k users to analyze socio-economic levels within an urban area in a Latin American city. Follow on studies have analyzed CDR data in conjunction with data from phone surveys using linear regression [14] or in conjunction with environmental data using Gaussian Process Regression [16]. The spatial resolution varied from sub-prefecture and lower administration level in Cote d'Ivoire [12], Democratic and Health Survey (DHS) clusters 1 in Rwanda [14], and regions and communes in Senegal [15,16].…”

Section: A Poverty Mapping Using Cdr Datamentioning

confidence: 99%

“…Follow on studies have analyzed CDR data in conjunction with data from phone surveys using linear regression [14] or in conjunction with environmental data using Gaussian Process Regression [16]. The spatial resolution varied from sub-prefecture and lower administration level in Cote d'Ivoire [12], Democratic and Health Survey (DHS) clusters 1 in Rwanda [14], and regions and communes in Senegal [15,16]. An annotated bibliography is provided in table I, summarizing details and results of the key research papers on poverty mapping using CDR data.…”

Section: A Poverty Mapping Using Cdr Datamentioning

confidence: 99%

“…Features for Model 1 were derived from transactions generated by 330k users acting over the sampling period and classed as living in Dar Es Salaam. They include a range of features identified as important for poverty prediction as identified by Pokhriyal et al [16]. The features can be broadly classified as falling under five different domains:…”

Section: Cdr Derived Featuresmentioning

confidence: 99%

See 1 more Smart Citation

The Unbanked and Poverty: Predicting area-level socio-economic vulnerability from M-Money transactions

Engelmann

Smith

Goulding

2018

2018 IEEE International Conference on Big Data (Big Data)

View full text Add to dashboard Cite

Emerging economies around the world are often characterized by governments and institutions struggling to keep key demographic data streams up to date. A demographic of interest particularly linked to social vulnerability is that of poverty and socioeconomic status. The combination of mass call detail records (CDR) data with machine learning has recently been proposed as a way to obtain this data without the expense required by traditional census and household survey methods. Based on a sample of 330k mobile phone subscribers resident in Dar es Salaam, Tanzania (7.6m M-Money records, 450.2m call and SMS event logs) this paper demonstrates the improvements that can be made via an alternate data stream: M-Money transaction records. An alternative to traditional banking services, particularly utilized by citizens unable to obtain a bank account, M-Money transactions provide a currently unexplored but potentially more powerful data set held by the same telecommunication companies. Comparing directly to CDR as used in prior work the results show that M-Money provides an increase in socio-demographic classification accuracy (average F1 score) from 65.9% (0.63) to 71.3% (0.7) at much finer-grained spatial regions than previously examined. Notably, the combined use of M-Money and CDR data only increases prediction accuracy (average F1 score) from 71.3% (0.7) to 72.3% (0.71), providing evidence that M-Money is informationally subsuming CDR data. The reasons for this and the importance/contributions of individual features are subsequently investigated.

show abstract

Combining disparate data sources for improved poverty prediction and mapping

Cited by 151 publications

References 42 publications

Mobile Phone Data for Urban Climate Change Adaptation: Reviewing Applications, Opportunities and Key Challenges

Mobile Phone Data for Urban Climate Change Adaptation: Reviewing Applications, Opportunities and Key Challenges

Credit Scoring Using Machine Learning by Combing Social Network Information: Evidence from Peer-to-Peer Lending

The Unbanked and Poverty: Predicting area-level socio-economic vulnerability from M-Money transactions

Contact Info

Product

Resources

About