Beyond Data Points: Regionalizing Crowdsourced Latency Measurements

Abstract

This paper develops a regionalization method for crowdsourced latency data, grouping geographic areas by similarity in internet performance. Applied to a major U.S. city, it reveals coherent regional patterns that cut across administrative boundaries and correlate with network infrastructure — a new lens for understanding internet inequity.

Cite

@inproceedings{sharma2025beyond,
  author = {Sharma, Taveesh and Schmitt, Paul and Bronzino, Francesco and Marwell, Nicole P. and Feamster, Nick},
  title = {Beyond Data Points: Regionalizing Crowdsourced Latency Measurements},
  booktitle = {ACM SIGMETRICS},
  year = {2025},
  url = {https://dl.acm.org/doi/10.1145/3700416}
}