pith. sign in

arxiv: 2604.16193 · v1 · submitted 2026-04-17 · ⚛️ physics.soc-ph

Correcting socioeconomic bias in mobile phone mobility estimates using multilevel regression and poststratification

Leo Ferres , Laetitia Gauvin This is my paper

Pith reviewed 2026-05-10 07:08 UTC · model grok-4.3

classification ⚛️ physics.soc-ph
keywords call detail recordshuman mobilitysocioeconomic biasmultilevel regressionpoststratificationradius of gyrationChilesampling correction
0
0 comments X

The pith

Multilevel regression and poststratification reduces biased mobile phone mobility estimates by 17% in Santiago.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Call detail records from a single mobile operator in Chile show users skewed toward higher socioeconomic groups, which inflates aggregate measures of how far people travel. The paper applies multilevel regression and poststratification to fit mobility patterns by socioeconomic status, gender, and location, then reweights predictions to match census demographics. This adjustment lowers the average radius of gyration compared with the raw data. A simpler model using only geographic information recovers much of the same correction, indicating that spatial sorting of socioeconomic groups can substitute for missing individual-level details.

Core claim

When call detail records from one Chilean carrier are treated as a population sample, the observed average radius of gyration is inflated because higher-mobility socioeconomic groups are overrepresented; fitting a Bayesian multilevel model that pools mobility estimates across comunas and then poststratifying the predictions to census population shares removes most of this distortion.

What carries the argument

Multilevel regression and poststratification (MRP): a Bayesian model that estimates individual mobility as a function of socioeconomic status, gender, and comuna, with partial pooling, followed by reweighting of predictions to match the full census demographic distribution.

If this is right

  • Single-operator call detail records systematically overestimate average travel distances when higher socioeconomic users are overrepresented.
  • Geographic poststratification alone can remove a substantial fraction of socioeconomic sampling bias even when individual user demographics are unknown.
  • Mobility studies can treat carrier data as non-representative samples and still produce population-level estimates by applying MRP.
  • Policy or research that relies on raw CDR aggregates for urban planning may need downward revision of typical travel ranges.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same correction could be tested on other mobility statistics such as number of unique locations visited or commuting flows.
  • In cities where census data exist only at coarse spatial scales, the geographic-only MRP variant offers a practical fallback.
  • Mobile operators could release pre-adjusted aggregates using this method to increase the reliability of downstream research.

Load-bearing premise

The relationship between socioeconomic status and mobility observed in the carrier data holds for the entire population, and the census provides an accurate, complete target distribution without other systematic differences.

What would settle it

Compare the MRP-adjusted average radius of gyration against the same metric computed from a large, probability-sampled travel survey or representative GPS panel for the same city and time period; a large remaining gap would indicate the correction is incomplete.

Figures

Figures reproduced from arXiv: 2604.16193 by Laetitia Gauvin, Leo Ferres.

Figure 1
Figure 1. Figure 1: CDR vs. census GSE proportions at the redcode level. [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Distribution of radius of gyration by GSE group. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Posterior distributions of fixed effects on log(ROG). Points show posterior means; [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Average radius of gyration by GSE: naive CDR average (grey) vs. MRP-corrected [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Naive vs. MRP-corrected average radius of gyration by comuna. [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Choropleth maps of the Santiago Metropolitan Region showing (left) naive CDR [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Radius of gyration by GSE group for three estimators: naive CDR average (red), [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
read the original abstract

Call detail records (CDR) from mobile phone networks are widely used to study human mobility however CDR data from a single mobile operator are inherently biased because the observed users do not mirror the population distribution. Using data from a major Chilean carrier in Santiago, we observe the user base is skewed by socioeconomic group, so aggregate metrics like radius of gyration are distorted by the population that is actually observed. To correct this sampling bias, we apply multilevel regression and poststratification (MRP), a method that is not yet standard for CDR-based mobility studies. We fit a Bayesian multilevel model for individual mobility using socioeconomic status, gender, and geography, with partial pooling across comunas, and then poststratify the predictions to match census demographics. This approach reduces the naive CDR estimate of average radius of gyration by about 17%. Importantly, a version of the model that uses only geographic information still captures much of the bias, showing that MRP can be useful even when the socioeconomic composition of users is not fully known, as long as spatial patterns of socioeconomic groups exist. This example demonstrates how MRP can provide a principled correction for non-representative CDR-derived mobility estimates, rather than treating the carrier sample as if it were a random population sample.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper applies multilevel regression and poststratification (MRP) to correct socioeconomic sampling bias in call detail records (CDR) from a single Chilean mobile operator in Santiago. A Bayesian multilevel model is fit to individual mobility (radius of gyration) using socioeconomic status, gender, and geography with partial pooling across comunas; predictions are then poststratified to census margins. The central claim is that this reduces the naive CDR estimate of average radius of gyration by ~17%, and that a geography-only variant recovers much of the correction even without full SES information.

Significance. If the correction is shown to be valid, the work would be useful for the many mobility studies that rely on non-representative CDR samples. The demonstration that spatial information alone can capture a large fraction of the bias is practically valuable when operator SES data are limited. The application of MRP (a standard survey-sampling tool) to CDR mobility is novel in this literature and, if accompanied by proper validation, could encourage more principled bias correction rather than treating carrier samples as population proxies.

major comments (3)
  1. [Abstract / Results] Abstract and Results: The headline 17% reduction in average radius of gyration is reported without uncertainty intervals, model diagnostics (e.g., posterior predictive checks, R-hat values, or effective sample sizes), or sensitivity tests to hyperparameter choices for partial pooling. This makes it impossible to judge whether the reported correction is statistically distinguishable from zero or robust to modeling decisions.
  2. [Methods] Methods: No external validation, hold-out test, or comparison against an independent mobility benchmark (e.g., travel survey, second operator, or GPS panel) is described to confirm that the poststratified E[mobility | SES, geography] is closer to the population truth than the raw CDR mean. The skeptic concern that within-stratum mobility may still differ between sampled and unsampled individuals due to unmeasured selection (age, plan type, differential travel distance) is therefore unaddressed.
  3. [Results] Results: The finding that the geography-only model captures “much of the bias” is presented as supportive, yet no quantitative decomposition or test distinguishes whether this reflects genuine spatial clustering of SES or simply reproduces the spatial distribution already present in the biased sample. This leaves open the possibility that the correction is partly circular with respect to the observed spatial sampling pattern.
minor comments (2)
  1. [Methods] The radius of gyration definition and the exact poststratification formula should be stated explicitly (including how comuna-level predictions are aggregated to the national or metropolitan margin) to allow direct replication.
  2. [Methods] Clarify the prior distributions and the precise form of the multilevel model (e.g., whether comuna-level intercepts are modeled as random effects with hyperpriors) so readers can assess the degree of shrinkage.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the detailed and constructive report. The comments highlight important areas for strengthening the presentation of uncertainty, validation, and interpretation of the geography-only results. We address each point below and indicate planned revisions.

read point-by-point responses

  1. Referee: [Abstract / Results] Abstract and Results: The headline 17% reduction in average radius of gyration is reported without uncertainty intervals, model diagnostics (e.g., posterior predictive checks, R-hat values, or effective sample sizes), or sensitivity tests to hyperparameter choices for partial pooling. This makes it impossible to judge whether the reported correction is statistically distinguishable from zero or robust to modeling decisions.

    Authors: We agree that the current presentation of the 17% reduction lacks sufficient statistical detail. In the revised manuscript we will report credible intervals for the poststratified mean radius of gyration and the percentage reduction, include standard model diagnostics (R-hat, effective sample size, and posterior predictive checks), and add a sensitivity analysis varying the partial-pooling hyperparameters. These additions will be placed in both the Results section and a new supplementary table. revision: yes

  2. Referee: [Methods] Methods: No external validation, hold-out test, or comparison against an independent mobility benchmark (e.g., travel survey, second operator, or GPS panel) is described to confirm that the poststratified E[mobility | SES, geography] is closer to the population truth than the raw CDR mean. The skeptic concern that within-stratum mobility may still differ between sampled and unsampled individuals due to unmeasured selection (age, plan type, differential travel distance) is therefore unaddressed.

    Authors: We acknowledge that external validation would strengthen the claim. Our dataset consists of CDR from a single operator and does not contain linked records from travel surveys or other operators that would permit a direct benchmark. We will expand the Discussion to explicitly state the unverifiable assumption that, conditional on the included covariates, mobility is exchangeable between sampled and unsampled individuals within strata, and we will discuss the implications of possible residual selection on age or plan type. revision: partial

  3. Referee: [Results] Results: The finding that the geography-only model captures “much of the bias” is presented as supportive, yet no quantitative decomposition or test distinguishes whether this reflects genuine spatial clustering of SES or simply reproduces the spatial distribution already present in the biased sample. This leaves open the possibility that the correction is partly circular with respect to the observed spatial sampling pattern.

    Authors: The geography-only model is fit on the sample but then poststratified to the independent census population margins by comuna. This adjustment is not circular because the target distribution is external. We will add a quantitative decomposition that isolates the contribution of the spatial poststratification step versus the SES adjustment, and we will include a supplementary comparison of the comuna-level sampling weights implied by the data versus the census to demonstrate that the correction moves the estimates toward the population spatial distribution rather than merely echoing the sample. revision: yes

standing simulated objections not resolved
  • External validation against an independent mobility benchmark (travel survey, second operator, or GPS panel), which is unavailable in the current single-operator dataset.

Circularity Check

0 steps flagged

No significant circularity; standard MRP application yields computed correction

full rationale

The paper fits a Bayesian multilevel regression on observed CDR mobility (as a function of SES, gender, and geography with partial pooling across comunas) then applies poststratification to census margins. The reported 17% reduction in radius of gyration is an output of this external statistical procedure, not an input parameter or quantity defined by the target metric itself. The geographic-only variant is presented as a robustness check, not a redefinition. No self-citations, uniqueness theorems, or ansatzes from prior author work are invoked to justify the central claim. The derivation is self-contained against the observed sample and census benchmarks; any concerns about residual confounding belong to validity rather than circularity.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard Bayesian multilevel modeling assumptions and the accuracy of external census data; no new entities are introduced.

free parameters (1)
  • hyperparameters for partial pooling across comunas
    Bayesian multilevel model requires estimation of variance components that control information sharing between neighborhoods.
axioms (2)
  • domain assumption Census demographics provide an accurate target distribution for poststratification.
    Poststratification step assumes the census counts match the true population composition.
  • domain assumption Individual mobility is a function of socioeconomic status, gender, and geography with partial pooling.
    Core modeling assumption stated in the abstract.

pith-pipeline@v0.9.0 · 5523 in / 1226 out tokens · 95444 ms · 2026-05-10T07:08:15.196436+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

  1. [1]

    ismtchile : Calculating Socio Material Territorial Index , 2023

    Martín Rosas Araya. ismtchile : Calculating Socio Material Territorial Index , 2023. URL https://CRAN.R-project.org/package=ismtchile. R package version 2.1.5. Proyecto Índice Socio Material Territorial. Proyecto Fondecyt DIMES, Centro de Investigación Urbano-Territorial, Observatorio de Ciudades, Universidad Católica de Chile

    work page 2023

  2. [2]

    and Lenormand, Maxime and Louail, Thomas and Menezes, Ronaldo and Ramasco, José J

    Hugo Barbosa, Marc Barthelemy, Gautam Ghoshal, Charlotte R. James, Maxime Lenormand, Thomas Louail, Ronaldo Menezes, Jos \'e J. Ramasco, Filippo Simini, and Marcello Tomasini. Human mobility: models and applications. Physics Reports, 734: 0 1--74, 2018. doi:10.1016/j.physrep.2018.01.001

    work page doi:10.1016/j.physrep.2018.01.001 2018

  3. [3]

    Mobile divides: gender, socioeconomic status, and mobile phone use in Rwanda

    Joshua Blumenstock and Nathan Eagle. Mobile divides: gender, socioeconomic status, and mobile phone use in Rwanda . In Proceedings of the 4th ACM/IEEE International Conference on Information and Communication Technologies and Development, 2010. doi:10.1145/2369220.2369225

    work page doi:10.1145/2369220.2369225 2010

  4. [4]

    Predicting poverty and wealth from mobile phone metadata

    Joshua Blumenstock, Gabriel Cadamuro, and Robert On. Predicting poverty and wealth from mobile phone metadata. Science, 350 0 (6264): 0 1073--1076, 2015. doi:10.1126/science.aac4420

    work page doi:10.1126/science.aac4420 2015

  5. [5]

    Buttice and Benjamin Highton

    Matthew K. Buttice and Benjamin Highton. How does multilevel regression and poststratification perform with conventional national surveys? Political Analysis, 21 0 (4): 0 449--467, 2013. doi:10.1093/pan/mpt017

    work page doi:10.1093/pan/mpt017 2013

  6. [6]

    A systematic machine learning approach to measure and assess biases in mobile phone population data, August 2025

    Carmen Cabrera and Francisco Rowe. A systematic machine learning approach to measure and assess biases in mobile phone population data, August 2025. URL http://arxiv.org/abs/2509.02603. arXiv:2509.02603 [stat]

    work page arXiv 2025

  7. [7]

    Tom \'a s Capretto, Camen Piho, Ravin Kumar, Jacob Westfall, Tal Yarkoni, and Osvaldo A. Martin. Bambi : a simple interface for fitting Bayesian linear models in Python . Journal of Statistical Software, 103 0 (15): 0 1--29, 2022. doi:10.18637/jss.v103.i15

    work page doi:10.18637/jss.v103.i15 2022

  8. [8]

    Buckee, and Amy Wesolowski

    Tiffany Chin, Bernardo Garc\'ia Cordova, Emily Laber, Yiquan Teng, Caroline O. Buckee, and Amy Wesolowski. Bias in mobility datasets drives divergence in modeled outbreak dynamics. Communications Medicine, 5: 0 45, 2025. doi:10.1038/s43856-025-00751-0

    work page doi:10.1038/s43856-025-00751-0 2025

  9. [9]

    Gaps in gender and socioeconomic mobility disparity studies

    Laetitia Gauvin. Gaps in gender and socioeconomic mobility disparity studies. Nature Computational Science, 4 0 (9): 0 633--635, 2024

    work page 2024

  10. [10]

    Gender gaps in urban mobility

    Laetitia Gauvin, Michele Tizzoni, Simone Piaggesi, Andrew Young, Natalia Adler, Stefaan Verhulst, Leo Ferres, and Ciro Cattuto. Gender gaps in urban mobility. Humanities and Social Sciences Communications, 7 0 (1): 0 1--13, 2020. doi:10.1057/s41599-020-0500-x

    work page doi:10.1057/s41599-020-0500-x 2020

  11. [11]

    Andrew Gelman and Thomas C. Little. Poststratification into many categories using hierarchical logistic regression. Survey Methodology, 33 0 (2): 0 127--135, 1997. URL https://sites.stat.columbia.edu/gelman/research/published/poststrat3.pdf

    work page 1997

  12. [12]

    Nature , volume =

    Marta C. Gonz \'a lez, C \'e sar A. Hidalgo, and Albert-L \'a szl \'o Barab \'a si. Understanding individual human mobility patterns. Nature, 453 0 (7196): 0 779--782, 2008. doi:10.1038/nature06958

    work page doi:10.1038/nature06958 2008

  13. [13]

    Estimating the effect of social inequalities on the mitigation of covid-19 across communities in santiago de chile

    Nicolò Gozzi, Michele Tizzoni, Matteo Chinazzi, Leo Ferres, Alessandro Vespignani, and Nicola Perra. Estimating the effect of social inequalities on the mitigation of covid-19 across communities in santiago de chile. Nature Communications, 12 0 (1), April 2021. ISSN 2041-1723. doi:10.1038/s41467-021-22601-6. URL http://dx.doi.org/10.1038/s41467-021-22601-6

    work page doi:10.1038/s41467-021-22601-6 2021

  14. [14]

    Modeling the Bias of Digital Data : An Approach to Combining Digital With Official Statistics to Estimate and Predict Migration Trends

    Yuan Hsiao, Lee Fiorio, Jonathan Wakefield, and Emilio Zagheni. Modeling the Bias of Digital Data : An Approach to Combining Digital With Official Statistics to Estimate and Predict Migration Trends . Sociological Methods & Research, 53 0 (4): 0 1905--1943, November 2024. ISSN 0049-1241. doi:10.1177/00491241221140144. URL https://doi.org/10.1177/00491241221140144

    work page doi:10.1177/00491241221140144 1905

  15. [15]

    Censo de poblaci \'o n y vivienda 2017, 2017

    Instituto Nacional de Estad \'i sticas . Censo de poblaci \'o n y vivienda 2017, 2017. Santiago, Chile

    work page 2017

  16. [16]

    Lax and Justin H

    Jeffrey R. Lax and Justin H. Phillips. How should we estimate public opinion in the states? American Journal of Political Science, 53 0 (1): 0 107--121, 2009. doi:10.1111/j.1540-5907.2008.00360.x

    work page doi:10.1111/j.1540-5907.2008.00360.x 2009

  17. [17]

    Cant \'u -Ros, Miguel Picornell, Ricardo Herranz, Juan Murillo Arias, Marc Barthelemy, Maxi San Miguel, and Jos \'e J

    Maxime Lenormand, Thomas Louail, Oliva G. Cant \'u -Ros, Miguel Picornell, Ricardo Herranz, Juan Murillo Arias, Marc Barthelemy, Maxi San Miguel, and Jos \'e J. Ramasco. Influence of sociodemographics on human mobility. Scientific Reports, 5: 0 10075, 2015. doi:10.1038/srep10075

    work page doi:10.1038/srep10075 2015

  18. [18]

    Understanding the bias of mobile location data across spatial scales and over time: a comprehensive analysis of safegraph data in the united states

    Zhenlong Li, Huan Ning, Fengrui Jing, and M Naser Lessani. Understanding the bias of mobile location data across spatial scales and over time: a comprehensive analysis of safegraph data in the united states. Plos one, 19 0 (1): 0 e0294430, 2024

    work page 2024

  19. [19]

    Nonrepresentativeness of human mobility data and its impact on modeling dynamics of the covid-19 pandemic: Systematic evaluation

    Chuchu Liu, Petter Holme, Sune Lehmann, Wenchuan Yang, and Xin Lu. Nonrepresentativeness of human mobility data and its impact on modeling dynamics of the covid-19 pandemic: Systematic evaluation. JMIR Formative Research, 8: 0 e55013, June 2024. ISSN 2561-326X. doi:10.2196/55013. URL http://dx.doi.org/10.2196/55013

    work page doi:10.2196/55013 2024

  20. [20]

    Beware Thy Bias : Scaling Mobile Phone Data to Measure Traffic Intensities

    Johan Meppelink, Jens Van Langen, Arno Siebes, and Marco Spruit. Beware Thy Bias : Scaling Mobile Phone Data to Measure Traffic Intensities . Sustainability, 12 0 (9): 0 3631, January 2020. ISSN 2071-1050. doi:10.3390/su12093631. URL https://www.mdpi.com/2071-1050/12/9/3631

    work page doi:10.3390/su12093631 2020

  21. [21]

    Returners and explorers dichotomy in human mobility

    Luca Pappalardo, Filippo Simini, Salvatore Rinzivillo, Dino Pedreschi, Fosca Giannotti, and Albert-L \'a szl \'o Barab \'a si. Returners and explorers dichotomy in human mobility. Nature Communications, 6: 0 8166, 2015. doi:10.1038/ncomms9166

    work page doi:10.1038/ncomms9166 2015

  22. [22]

    Evaluation of home detection algorithms on mobile phone data using individual-level ground truth

    Luca Pappalardo, Leo Ferres, Manuel Sacasa, Ciro Cattuto, and Loreto Bravo. Evaluation of home detection algorithms on mobile phone data using individual-level ground truth. EPJ Data Science, 10 0 (1): 0 29, 2021. doi:10.1140/epjds/s13688-021-00284-9

    work page doi:10.1140/epjds/s13688-021-00284-9 2021

  23. [23]

    Park, Andrew Gelman, and Joseph Bafumi

    David K. Park, Andrew Gelman, and Joseph Bafumi. Bayesian multilevel estimation with poststratification: state-level estimates from national polls. Political Analysis, 12 0 (4): 0 375--385, 2004. doi:10.1093/pan/mph024

    work page doi:10.1093/pan/mph024 2004

  24. [24]

    The ABCDE of Big Data : Assessing Biases in Call - Detail Records for Development Estimates

    Gabriel Pestre, Emmanuel Letouzé, and Emilio Zagheni. The ABCDE of Big Data : Assessing Biases in Call - Detail Records for Development Estimates . The World Bank Economic Review, 34 0 (Supplement\_1): 0 S89--S97, February 2020. ISSN 0258-6770, 1564-698X. doi:10.1093/wber/lhz039. URL https://academic.oup.com/wber/article/34/Supplement_1/S89/5650800

    work page doi:10.1093/wber/lhz039 2020

  25. [25]

    Smartphone ownership and internet usage continues to climb in emerging economies, 2016

    Pew Research Center . Smartphone ownership and internet usage continues to climb in emerging economies, 2016. URL https://www.pewresearch.org/global/2016/02/22/smartphone-ownership-and-internet-usage-continues-to-climb-in-emerging-economies/

    work page 2016

  26. [26]

    Beyond the ``single-operator, CDR -only'' paradigm: an interoperable framework for mobile phone network data analyses and population density estimation

    Fabio Ricciato, Peter Widhalm, Francesco Pantisano, and Max Craglia. Beyond the ``single-operator, CDR -only'' paradigm: an interoperable framework for mobile phone network data analyses and population density estimation. Pervasive and Mobile Computing, 35: 0 65--82, 2017. doi:10.1016/j.pmcj.2016.04.009

    work page doi:10.1016/j.pmcj.2016.04.009 2017

  27. [27]

    Wiecki, and Christopher Fonnesbeck

    John Salvatier, Thomas V. Wiecki, and Christopher Fonnesbeck. Probabilistic programming in Python using PyMC3 . PeerJ Computer Science, 2: 0 e55, 2016. doi:10.7717/peerj-cs.55

    work page doi:10.7717/peerj-cs.55 2016

  28. [28]

    Sanchez, Hamish Gibbs, Takahiro Yabe, Daniel T

    Sarah A. Sanchez, Hamish Gibbs, Takahiro Yabe, Daniel T. O'Brien, and Esteban Moro. Correcting temporal bias in mobility data using time-use surveys, January 2026. URL https://arxiv.org/abs/2601.22330v1

    work page arXiv 2026

  29. [29]

    Sector Telecomunicaciones: Cierre 2016

    Subsecretaría de Telecomunicaciones de Chile . Sector Telecomunicaciones: Cierre 2016 . https://www.subtel.gob.cl/wp-content/uploads/2016/12/PPT_Series_DICIEMBRE_2016_V3.pdf, December 2016. Accessed: 2026-04-15

    work page 2016

  30. [30]

    Measuring under-5 mortality and fertility through mobile phone surveys: an assessment of selection bias in 34 low-income and middle-income countries

    David A Sánchez-Páez, Bruno Masquelier, Ashira Menashe-Oren, Ololade Julius Baruwa, and Georges Reniers. Measuring under-5 mortality and fertility through mobile phone surveys: an assessment of selection bias in 34 low-income and middle-income countries. BMJ Open, 13 0 (11): 0 e071791, November 2023. ISSN 2044-6055, 2044-6055. doi:10.1136/bmjopen-2023-071...

    work page doi:10.1136/bmjopen-2023-071791 2023

  31. [31]

    Schneider, Vincent Blondel, Zbigniew Smoreda, Marta C

    Michele Tizzoni, Paolo Bajardi, Adeline Decuyper, Gary King King, Christian M. Schneider, Vincent Blondel, Zbigniew Smoreda, Marta C. Gonz \'a lez, and Vittoria Colizza. On the use of human mobility proxies for modeling epidemics. PLoS Computational Biology, 10 0 (7): 0 e1003716, 2014. doi:10.1371/journal.pcbi.1003716

    work page doi:10.1371/journal.pcbi.1003716 2014

  32. [32]

    Assessing the Quality of Home Detection from Mobile Phone Data for Official Statistics , volume=

    Maarten Vanhoof, Fernando Reis, Thomas Ploetz, and Zbigniew Smoreda. Assessing the quality of home detection from mobile phone data for official statistics. Journal of Official Statistics, 34 0 (4): 0 935--960, 2018. doi:10.2478/jos-2018-0046

    work page doi:10.2478/jos-2018-0046 2018

  33. [33]

    and Gelman, A

    Aki Vehtari, Andrew Gelman, and Jonah Gabry. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC . Statistics and Computing, 27 0 (5): 0 1413--1432, 2017. doi:10.1007/s11222-016-9696-4

    work page doi:10.1007/s11222-016-9696-4 2017

  34. [34]

    Forecasting elections with non-representative polls

    Wei Wang, David Rothschild, Sharad Goel, and Andrew Gelman. Forecasting elections with non-representative polls. International Journal of Forecasting, 31 0 (3): 0 980--991, July 2015. ISSN 0169-2070. doi:10.1016/j.ijforecast.2014.06.001. URL https://www.sciencedirect.com/science/article/pii/S0169207014000879

    work page doi:10.1016/j.ijforecast.2014.06.001 2015

  35. [35]

    Tatem, David L

    Amy Wesolowski, Nathan Eagle, Andrew J. Tatem, David L. Smith, Abdisalan M. Noor, Robert W. Snow, and Caroline O. Buckee. Quantifying the impact of human mobility on malaria. Science, 338 0 (6104): 0 267--270, 2012. doi:10.1126/science.1223467

    work page doi:10.1126/science.1223467 2012

  36. [37]

    Noor, Robert W

    Amy Wesolowski, Nathan Eagle, Abdisalan M. Noor, Robert W. Snow, and Caroline O. Buckee. The impact of biases in mobile phone ownership on estimates of human mobility. Journal of The Royal Society Interface, 10 0 (81): 0 20120986, April 2013 b . ISSN 1742-5689. doi:10.1098/rsif.2012.0986. URL https://doi.org/10.1098/rsif.2012.0986

    work page doi:10.1098/rsif.2012.0986 2013

  37. [38]

    Understanding the bias of call detail records in human mobility research

    Ziliang Zhao, Shih-Lung Shaw, Yang Xu, Feng Lu, Jie Chen, and Ling Yin. Understanding the bias of call detail records in human mobility research. International Journal of Geographical Information Science, 30 0 (9): 0 1738--1762, September 2016. ISSN 1365-8816. doi:10.1080/13658816.2015.1137298. URL https://doi.org/10.1080/13658816.2015.1137298. \_eprint: ...

    work page doi:10.1080/13658816.2015.1137298 2016