Published Research

Social Determinants of Health

American Journal of Epidemiology · 2026

@article{yan2026sdoh,
  title   = {Revisiting the social determinants of health with explainable AI: a cross-country perspective},
  author  = {Yan, Jiani},
  journal = {American Journal of Epidemiology},
  volume  = {195},
  number  = {3},
  pages   = {681--688},
  year    = {2026},
  doi     = {10.1093/aje/kwaf205}
}

Revisiting the social determinants of health with explainable AI: a cross-country perspective on how social factors shape mortality and morbidity outcomes.

  • Among people aged 50 and over, age and gender are the consistent predictors of mortality across every context studied.
  • Demographic and socioeconomic factors are the most important social domains for predicting death in all contexts.
  • The importance of other social factors is age-dependent and context-dependent.

Prediction Limits

Nature Computational Science · 2025

@article{yan2025limits,
  title   = {On the unknowable limits to prediction},
  author  = {Yan, Jiani and Rahal, Charles},
  journal = {Nature Computational Science},
  volume  = {5},
  number  = {3},
  pages   = {188--190},
  year    = {2025},
  doi     = {10.1038/s43588-025-00776-y}
}

On the unknowable limits to prediction — a perspective on what computational social science can and cannot forecast.

Digital Gender Gaps

PNAS · 2025

@article{breen2025dgg,
  title   = {Mapping subnational gender gaps in internet and mobile adoption using
             social media data},
  author  = {Breen, Casey F. and Fatehkia, Masoomali and Yan, Jiani and Zhao, Xinyi
             and Leasure, Douglas R. and Weber, Ingmar and Kashyap, Ridhi},
  journal = {Proceedings of the National Academy of Sciences},
  volume  = {122},
  number  = {42},
  pages   = {e2416624122},
  year    = {2025},
  doi     = {10.1073/pnas.2416624122}
}

Mapping subnational gender gaps in internet and mobile adoption using social media advertising data, with a live global dashboard.

Population Displacement in Ukraine

Population and Development Review · 2023

@article{leasure2023ukraine,
  title   = {Nowcasting Daily Population Displacement in Ukraine through Social
             Media Advertising Data},
  author  = {Leasure, Douglas R. and Kashyap, Ridhi and Rampazzo, Francesco and
             Dooley, Claire A. and Elbers, Benjamin and Bondarenko, Maksym and
             Verhagen, Mark and Frey, Arun and Yan, Jiani and Akimova, Evelina T.
             and Fatehkia, Masoomali and Trigwell, Robert and Tatem, Andrew J.
             and Weber, Ingmar and Mills, Melinda C.},
  journal = {Population and Development Review},
  volume  = {49},
  number  = {2},
  pages   = {231--254},
  year    = {2023},
  doi     = {10.1111/padr.12558}
}

Nowcasting daily population displacement through social media advertising data.

Scientific Software

RobustiPy

Patterns · 2026

@article{valdenegro2026robustipy,
  title   = {RobustiPy: An efficient next-generation multiversal library with model
             selection, averaging, resampling, and explainable AI},
  author  = {Valdenegro, Daniel and Yan, Jiani and Dai, Duiyi and Rahal, Charles},
  journal = {Patterns},
  volume  = {7},
  number  = {8},
  pages   = {101609},
  year    = {2026},
  doi     = {10.1016/j.patter.2026.101609}
}

An efficient next-generation multiversal library with model selection, averaging, resampling and explainable AI — enumerate every defensible specification, then assess robustness and interpretability.

GWAS Diversity Monitor

Live dashboard

A live dashboard monitoring diversity in genome-wide association studies, tracking representation across ancestry groups in genetic research.

Digital Gender Gaps Web App

Zenodo · 2023

@misc{leasure2023dggapp,
  title     = {Digital Gender Gaps Web Application, v1.0.0},
  author    = {Leasure, Douglas R. and Yan, Jiani and Bondarenko, Maksym and
               Kerr, David and Fatehkia, Masoomali and Weber, Ingmar and
               Kashyap, Ridhi},
  publisher = {Zenodo},
  year      = {2023},
  doi       = {10.5281/zenodo.7897491}
}

The web application behind the Digital Gender Gaps project, archived on Zenodo and developed openly on GitHub.

In Progress

UK Biobank: Disease Profiles

Discovering prevalent chronic disease profiles with advanced clustering methods in UK Biobank.

Multimorbidity Trajectories

An item response theory approach to modelling multimorbidity complexity trajectories and the role of social determinants in UK Biobank.

On the Dangers of a Single Seed

How the choice of a single random seed propagates through machine learning pipelines, and why that is a danger in applied scientific research.

Nowcasting Digital Gender Gaps

Nowcasting global digital gender gaps using social media data, extending the subnational mapping work to near real-time estimates.