Spatial Analysis

Overview

Understanding Internet performance requires analyzing data at fine geographic granularities. Our spatial analysis research develops methods for measuring, mapping, and modeling Internet performance variations across neighborhoods, census tracts, and even individual blocks.

Research Areas

Hyperlocal Performance Variability

We study how Internet performance varies at hyperlocal scales—within neighborhoods and across small geographic areas. This research reveals that aggregate statistics often mask significant local disparities that affect residents’ actual experience.

IP Geolocation

Accurate geolocation of Internet measurements is critical for spatial analysis. We developed GPS-based methods for geolocating consumer IP addresses, enabling more precise mapping of Internet performance than traditional IP geolocation databases.

Spatial Sampling Methods

Effective broadband mapping requires careful attention to sampling methodology. We investigate how to design measurement studies that capture meaningful spatial variation while remaining feasible to deploy at scale.

Multi-Level Modeling

Internet performance is influenced by factors at multiple geographic scales—from individual households to neighborhoods to metropolitan regions. We apply multi-level statistical models to understand how these factors interact across scales.

Key Findings

• Internet performance can vary significantly even within small geographic areas
• Traditional IP geolocation methods introduce substantial errors in performance mapping
• Spatial sampling design critically affects the conclusions that can be drawn from measurement studies

Publications

A First Look at the Spatial and Temporal Variability of Internet Performance Data in Hyperlocal Geographies Taveesh Sharma, Jonatas Marques, Nick Feamster, Nicole Marwell TPRC, 2023

GPS-Based Geolocation of Consumer IP Addresses Jamie Saxon, Nick Feamster Passive and Active Measurement Conference (PAM), 2022

Collaborators

This research involves collaboration with researchers at the University of Chicago and USC, combining expertise in computer science, sociology, and urban policy.