Analyzing the Effect of Time in Migration Measurement Using Georeferenced Digital Trace Data.

Georeferenced digital trace data offer unprecedented flexibility in migration estimation. Because of their high temporal granularity, many migration estimates can be generated from the same data set by changing the definition parameters. Yet despite the growing application of digital trace data to migration research, strategies for taking advantage of their temporal granularity remain largely underdeveloped. In this paper, we provide a general framework for converting digital trace data into estimates of migration transitions and for systematically analyzing their variation along a quasi-continuous time scale, analogous to a survival function. From migration theory, we develop two simple hypotheses regarding how we expect our estimated migration transition functions to behave. We then test our hypotheses on simulated data and empirical data from three platforms in two internal migration contexts: geotagged Tweets and Gowalla check-ins in the United States, and cell-phone call detail records in Senegal. Our results demonstrate the need for evaluating the internal consistency of migration estimates derived from digital trace data before using them in substantive research. At the same time, however, common patterns across our three empirical data sets point to an emergent research agenda using digital trace data to study the specific functional relationship between estimates of migration and time and how this relationship varies by geography and population characteristics.

Accurate intercensal estimates of energy access to track Sustainable Development Goal 7.

Intercensal estimates of access to electricity and clean cooking fuels at policy planning microregions in a country are essential for understanding their evolution and tracking progress towards Sustainable Development Goals (SDG) 7. Surveys are prohibitively expensive to get such intercensal microestimates. Existing works, mainly, focus on electrification rates, make predictions at the coarse spatial granularity, and generalize poorly to intercensal periods. Limited works focus on estimating clean cooking fuel access, which is one of the crucial indicators for measuring progress towards SDG 7. We propose a novel spatio-temporal multi-target Bayesian regression model that provides accurate intercensal microestimates for household electrification and clean cooking fuel access by combining multiple types of earth-observation data, census, and surveys. Our model's estimates are produced for Senegal for 2020 at policy planning microregions, and they explain 77% and 86% of variation in regional aggregates for electrification and clean fuels, respectively, when validated against the most recent survey. The diagnostic nature of our microestimates reveals a slow evolution and significant lack of clean cooking fuel access in both urban and rural areas in Senegal. It underscores the challenge of expanding energy access even in urban areas owing to their rapid population growth. Owing to the timeliness and accuracy of our microestimates, they can help plan interventions by local governments or track the attainment of SDGs when no ground-truth data are available. Supplementary Information: The online version contains supplementary material available at 10.1140/epjds/s13688-022-00371-5.

Using Facebook advertising data to describe the socio-economic situation of Syrian refugees in Lebanon.

While the fighting in the Syrian civil war has mostly stopped, an estimated 5.6 million Syrians remain living in neighboring countries. Of these, an estimated 1.5 million are sheltering in Lebanon. Ongoing efforts by organizations such as UNHCR to support the refugee population are often ineffective in reaching those most in need. According to UNHCR's 2019 Vulnerability Assessment of Syrian Refugees Report (VASyR), only 44% of the Syrian refugee families eligible for multipurpose cash assistance were provided with help, as the others were not captured in the data. In this project, we are investigating the use of non-traditional data, derived from Facebook advertising data, for population level vulnerability assessment. In a nutshell, Facebook provides advertisers with an estimate of how many of its users match certain targeting criteria, e.g., how many Facebook users currently living in Beirut are "living abroad," aged 18-34, speak Arabic, and primarily use an iOS device. We evaluate the use of such audience estimates to describe the spatial variation in the socioeconomic situation of Syrian refugees across Lebanon. Using data from VASyR as ground truth, we find that iOS device usage explains 90% of the out-of-sample variance in poverty across the Lebanese governorates. However, evaluating predictions at a smaller spatial resolution also indicate limits related to sparsity, as Facebook, for privacy reasons, does not provide audience estimates for fewer than 1,000 users. Furthermore, comparing the population distribution by age and gender of Facebook users with that of the Syrian refugees from VASyR suggests an under-representation of Syrian women on the social media platform. This work adds to growing body of literature demonstrating the value of anonymous and aggregate Facebook advertising data for analysing large-scale humanitarian crises and migration events.

Socio-economic, built environment, and mobility conditions associated with crime: a study of multiple cities.

Nowadays, 23% of the world population lives in multi-million cities. In these metropolises, criminal activity is much higher and violent than in either small cities or rural areas. Thus, understanding what factors influence urban crime in big cities is a pressing need. Seminal studies analyse crime records through historical panel data or analysis of historical patterns combined with ecological factor and exploratory mapping. More recently, machine learning methods have provided informed crime prediction over time. However, previous studies have focused on a single city at a time, considering only a limited number of factors (such as socio-economical characteristics) and often at large in a single city. Hence, our understanding of the factors influencing crime across cultures and cities is very limited. Here we propose a Bayesian model to explore how violent and property crimes are related not only to socio-economic factors but also to the built environmental (e.g. land use) and mobility characteristics of neighbourhoods. To that end, we analyse crime at small areas and integrate multiple open data sources with mobile phone traces to compare how the different factors correlate with crime in diverse cities, namely Boston, Bogotá, Los Angeles and Chicago. We find that the combined use of socio-economic conditions, mobility information and physical characteristics of the neighbourhood effectively explain the emergence of crime, and improve the performance of the traditional approaches. However, we show that the socio-ecological factors of neighbourhoods relate to crime very differently from one city to another. Thus there is clearly no "one fits all" model.

Moves on the Street: Classifying Crime Hotspots Using Aggregated Anonymized Data on People Dynamics.

The wealth of information provided by real-time streams of data has paved the way for life-changing technological advancements, improving the quality of life of people in many ways, from facilitating knowledge exchange to self-understanding and self-monitoring. Moreover, the analysis of anonymized and aggregated large-scale human behavioral data offers new possibilities to understand global patterns of human behavior and helps decision makers tackle problems of societal importance. In this article, we highlight the potential societal benefits derived from big data applications with a focus on citizen safety and crime prevention. First, we introduce the emergent new research area of big data for social good. Next, we detail a case study tackling the problem of crime hotspot classification, that is, the classification of which areas in a city are more likely to witness crimes based on past data. In the proposed approach we use demographic information along with human mobility characteristics as derived from anonymized and aggregated mobile network data. The hypothesis that aggregated human behavioral data captured from the mobile network infrastructure, in combination with basic demographic information, can be used to predict crime is supported by our findings. Our models, built on and evaluated against real crime data from London, obtain accuracy of almost 70% when classifying whether a specific area in the city will be a crime hotspot or not in the following month.