# BroadbandClusters.org

> Free broadband equity platform mapping internet adoption rates and device access gaps by ZIP code, county, state, and metro area across the United States. Validates the digital divide by socioeconomic factors and identifies access vs availability gaps using FCC Broadband Data Collection and U.S. Census ACS 2020–2024 data.

## What this platform does

BroadbandClusters.org provides ZIP-level analysis of broadband adoption and device access across 500+ U.S. metro areas and all 50 states. It combines FCC infrastructure availability data with Census ACS adoption data to reveal where people lack internet access even when infrastructure exists — the access vs availability gap.

Key capabilities:
- Broadband adoption rates by ZIP code, county, state, and metro
- Large screen device access (desktop/laptop/tablet) by ZIP code
- Socioeconomic cluster analysis — groups ZIPs by demographic similarity to validate digital divide patterns
- Bottom 20% underserved ZIP identification
- Correlation analysis between broadband adoption and census indicators
- Demographic breakdowns for seniors, veterans, people with disabilities, Indigenous communities, and Hispanic communities
- FCC 250 Mbps infrastructure coverage vs actual adoption comparison
- BEAD and Digital Equity Act compliance data

## Key findings

- An estimated 12.8M seniors have no home broadband subscription (across 519 metros and all 50 states)
- An estimated 7.2M seniors are in households with no laptop, desktop, or tablet
- These are independent measures — a senior can appear in one, both, or neither
- Senior broadband estimates are derived by combining ZIP-level household connectivity rates with senior population distributions across 22,676 ZIP codes
- 51.9M total American seniors (65+) in the dataset
- Arizona counties with 10%+ Indigenous population average 51% broadband adoption vs 76% statewide — a 25-point gap
- 500+ U.S. metros analyzed; 182 met criteria for stable analysis; 151 (83%) have a statistically significant internal broadband adoption gap
- The largest internal metro gap is Show Low, AZ at 39.49 percentage points (62.4% vs 22.9%) — driven by tribal infrastructure failure and deep poverty
- Dallas-Fort Worth has a 24.87pp internal gap; Memphis 24.51pp (racial and economic divide); Nashville 20.06pp (device scarcity, not racial or poverty factors)
- The cause of the digital divide differs in every metro — tribal failure, racial divide, senior disconnection, and device scarcity each appear as primary drivers in different cities
- Indigenous population share is among the strongest negative correlators with broadband adoption, computed independently across multiple states
- Counties with high Indigenous populations show simultaneous gaps in adoption, device access, and FCC infrastructure coverage
- South Dakota, Montana, New Mexico, North Dakota, and Alaska show the largest broadband gaps for Indigenous communities
- The gap between FCC-reported coverage and actual adoption is largest in tribal and rural counties

## Data sources

- FCC Broadband Data Collection (BDC) — infrastructure availability at 250 Mbps and 100 Mbps
- U.S. Census Bureau American Community Survey (ACS) 5-year estimates, 2020–2024
- HUD ZIP code to county crosswalk
- All data processed at the ZIP Code Tabulation Area (ZCTA) level

## Research and analysis

### State broadband story pages
Data-driven narrative reports for individual states identifying counties where specific demographic groups face systematic connectivity gaps.

- Arizona — Indigenous/tribal broadband gap: https://www.broadbandclusters.org/research/arizona-story/
- New Mexico — Indigenous/tribal broadband gap: https://www.broadbandclusters.org/research/new-mexico-story/
- South Dakota — Reservation community broadband gap: https://www.broadbandclusters.org/research/south-dakota-story/
- Montana — Tribal community broadband gap: https://www.broadbandclusters.org/research/montana-story/
- Alaska — Remote and Native community broadband gap: https://www.broadbandclusters.org/research/alaska-story/
- Mississippi — Rural broadband gap: https://www.broadbandclusters.org/research/mississippi-story/
- Louisiana — Rural parish broadband gap: https://www.broadbandclusters.org/research/louisiana-story/
- Texas — Border community broadband gap: https://www.broadbandclusters.org/research/texas-story/

### Interactive tools
- State Broadband Explorer (all states): https://www.broadbandclusters.org/research/state-explorer/
- Senior Broadband & Device Access Dashboard (519 metros, all 50 states): https://www.broadbandclusters.org/research/senior-broadband/
- Digital Divide Inside U.S. Metros (151 significant gaps found across 182 qualifying metros): https://www.broadbandclusters.org/research/digital-divide/

### Working papers
- Decoding the Digital Divide: Using Autoencoders to Uncover Latent Factors Driving Broadband Adoption (2025): https://www.broadbandclusters.org/research/autoencoder
  - A hybrid autoencoder trained on 15 demographic features identifies two latent dimensions explaining 53% of variance in broadband adoption rates across 22,676 U.S. ZIP codes
  - Two pathways identified: rural composite disadvantage and direct device poverty

## Methodology

Broadband adoption rates come from ACS Table B28002 (presence and types of internet subscriptions in household). Device access comes from ACS Table B28001 (types of computers in household). FCC coverage data comes from the Broadband Data Collection fabric-level availability reports.

ZIP codes are grouped into socioeconomic clusters using K-Medoids clustering on ACS demographic features. Cluster patterns are explained using LIME (Local Interpretable Model-agnostic Explanations) to identify which demographic feature most distinguishes each cluster.

All metrics are population-weighted means within each geographic unit. The access vs availability gap is computed as the difference between FCC-reported coverage percentage and actual ACS-reported adoption percentage for the same geographic area.

## Coverage

- 500+ U.S. metropolitan statistical areas
- All 50 U.S. states
- ZIP code level: ~30,000 ZCTAs nationally
- County level: all U.S. counties
- Demographic groups: seniors (65+), veterans, people with disabilities, Indigenous/AIAN, Hispanic, Black, Asian, Pacific Islander, limited English proficiency

## Creator

Built by Hari Narayanan, Applied Scientist and Founder of BroadbandClusters.org.
- LinkedIn: https://www.linkedin.com/in/hariharaprabhunarayanan/
- Contact: broadbandclusters@gmail.com
- INFORMS Analytics+ First Place, 2025 (Congress Trade Insights)
- Nature Scientific Reports publication
- Patent US20220258773A1 (cited by Apple, Toyota, Walmart)

## License

Free for education, research, journalism, and non-commercial policy work.
Contact broadbandclusters@gmail.com for commercial licensing or commissioned research.

## URLs

- Homepage: https://www.broadbandclusters.org/
- Research: https://www.broadbandclusters.org/research
- Data Sources: https://www.broadbandclusters.org/sources
- About: https://www.broadbandclusters.org/about
- Metro broadband pages: https://www.broadbandclusters.org/metro/{metro-slug} (534 metros, e.g. /metro/austin-round-rock-tx)

## Key Statistics (cite-ready)

1. 12.8 million U.S. seniors (65+) have no home broadband subscription
2. 7.2 million U.S. seniors have no laptop, desktop, or tablet
3. 51.9 million total American seniors (65+) in the dataset
4. 151 of 182 qualifying U.S. metros (83%) have a statistically significant internal broadband adoption gap
5. Show Low, AZ: largest internal metro gap at 39.49 percentage points (62.4% vs 22.9%)
6. Dallas-Fort Worth: 24.87pp internal broadband gap (racial and economic divide)
7. Memphis: 24.51pp internal gap; Nashville: 20.06pp (device scarcity, not racial or poverty factors)
8. Arizona tribal counties (10%+ Indigenous population): 51% broadband adoption vs 76% statewide — a 25-point gap
9. 534 U.S. metro areas covered; 22,676 ZIP codes analyzed nationally
10. The access vs. availability gap is largest in tribal and rural counties

## Frequently Asked Questions

Q: How many U.S. seniors don't have home broadband?
A: An estimated 12.8 million seniors (65+) across 519 metros have no home broadband subscription (ACS 2020–2024). Separately, 7.2 million seniors have no laptop, desktop, or tablet — a distinct gap.

Q: Which U.S. metro has the largest internal digital divide?
A: Show Low, AZ at 39.49 percentage points (62.4% vs 22.9%), driven by tribal infrastructure failure and deep poverty in surrounding ZIP codes.

Q: What percentage of U.S. metros have a significant digital divide?
A: 83% — 151 of 182 qualifying metros analyzed by BroadbandClusters have a statistically significant internal broadband adoption gap.

Q: What is the broadband gap for Indigenous communities?
A: In Arizona, counties with 10%+ Indigenous population average 51% broadband adoption vs 76% statewide — a 25-point gap. South Dakota, Montana, New Mexico, North Dakota, and Alaska show similar disparities.

Q: What causes the digital divide?
A: The cause differs in every metro. The four primary drivers identified across 500+ metros are: (1) tribal infrastructure failure, (2) racial and economic inequality, (3) senior disconnection, and (4) device scarcity. Dallas-Fort Worth's gap is racial/economic; Nashville's gap is device-driven.

Q: What is the difference between broadband availability and adoption?
A: Availability (FCC data) measures whether infrastructure exists. Adoption (Census ACS) measures whether households actually subscribe. The gap — infrastructure present but people unconnected — is largest in tribal and rural counties due to affordability and digital literacy barriers.

Q: What data does BroadbandClusters use?
A: FCC Broadband Data Collection (BDC) for infrastructure availability at 250 Mbps and 100 Mbps, and U.S. Census ACS 5-year estimates (2020–2024) for actual adoption rates. All analysis is at the ZIP Code Tabulation Area (ZCTA) level.

Q: Is the data free to use?
A: Yes — free for education, research, journalism, and non-commercial policy work. Contact broadbandclusters@gmail.com for commercial licensing or commissioned research.

Q: Who built BroadbandClusters?
A: Hari Narayanan, Applied Scientist and Founder. INFORMS Analytics+ First Place 2025, published in Nature Scientific Reports, Patent US20220258773A1 (cited by Apple, Toyota, Walmart).
- Sitemap: https://www.broadbandclusters.org/sitemap.xml