RURAL AREAS ARE OVERREPRESENTED IN UNSERVED AREAS AND UNDERESTIMATED IN STATISTICS

Extending Broadband to Underserved Areas

Olivier Leroy & Drew Rosebush

Center On Rural Innovation

2023-10-18

Our slides are online:

Slides realized in Quarto hosted in GitHub

Data stack: SQL (PostgreSQL), R with DBI, targets, ggplot2 and “in house” packages

https://ruralinnovation.github.io/conf_URISA_2023/#/title-slide

Who are we?

Olivier 🇨🇵 Senior Data Engineer

Drew 🇺🇸 Lead Data Scientist

Center On Rural Innovation (CORI)

Advancing economic prosperity in rural America through the creation of inclusive tech economy ecosystem that support scalable entrepreneurship and tech job creation

Learn more at: https://ruralinnovation.us/

Defining Rural

“Rural” refers to the “nonmetro plus” definition: includes all nonmetro counties and all tracts classified as RUCA 4 or higher.

USDA’s Rural-Urban Community Area (RUCA) codes categorize rural areas by population density, urbanization, and daily commuting.

More nuanced rural definition, classifying rural places along a continuum.

Combining definitions: better encapsulates places that are rural in character.

Historical and new data context

F477 NBM
US Census Boundaries 2010 2020
Type of recording self declarative self declarative
Granularity Census block Locations
Services Mobile/Fixed Mobile/Fixed
Releases 2014-2021 2022 - Ongoing

F477: “at least one” rule

NBM’s BEAD definitions for Unserved area: an area in which not less than 80 percent of broadband-serviceable locations are unserved locations.

What’s on the table

Broadband Equity Access and Deployment (BEAD) Program provides $42.45B to expand high-speed internet access by planning, infrastructure deployment, and adoption programs

Minimum BEAD allocations range from $25M (US territories) to $100M (states and Puerto Rico)

Remaining BEAD allocations vary across states depending on the number and concentration of unserved locations

BEAD Priorities

Coverage

Primary goal is to extend broadband internet service (≥ 25Mbps/3Mbps) to all

Technology

Prioritizes projects designed to provide fiber connectivity directly to the end user

Affordability

Prioritizes projects that aim at improving affordability of service

Workflow: From FCC to our products

%%{init: {'theme': 'base', 'themeVariables': { 'fontSize': '20px', 'fontFamily': 'Bitter'}}}%%
flowchart LR
    A1[FCC Website] --> B1[Server]
    A2[TIGER Census block] --> C2
    subgraph Ingestion
    direction TB
    B1--> B2[Data Wrangling]
    B2--> B3[Populate source DB]
    end
    subgraph Transform
    direction TB
    C1[Remove Sat.]--> C2[Count Services]
    C2 --> C3[Blocks DB]
    end
    subgraph Serving
    direction TB
    D1[Prod. Server] --> D2[Analytic Tables]
    end
    Ingestion --> Transform
    C3 --> D1
    D1--> E1[Tools: <Br> Public CH <BR> Internal BCAT]
    D2--> E1
    D2 --> E4[Data team <BR> Broadband team]
    D2 --> E5[URISA]
    click E1 "https://broadband-risk.ruralinnovation.us/" _parent
    click E5 "https://ruralinnovation.github.io/conf_URISA_2023/#/workflow-from-fcc-to-our-products" _parent
    style Ingestion fill:#fff
    style Transform fill:#fff
    style Serving fill:#fff
  • Two releases per year (but 7 versions of Dec. 2022 releases)

  • 440 files for 11GB zipped (Fixed)

  • This presentation uses FCC staff estimates

Results: At the US level

Rural areas Overrepresented in poor quality broadband!

Rural Areas underestimated in statistics

Picking a 50% threshold can help close the gap!

Conclusions

While nonrural access to broadband has exceeded 94% since 2014, rural broadband access stood at just 65% in 2014 and only reached 92% in 2020.

  • Overall great improvement in data quality and better understanding on what is happening on the ground (is it enough?)

  • Rural areas are still overrepresented in poor quality broadband

  • Rural areas are systematically underestimated by data definitions

Next steps

What are we measuring? An improvement on coverage, an improvement on data quality or both?

  1. Getting better at counting people (use Housing Units, Microsoft building footprint, ACS)

  2. Getting better at estimating locations (optimize the cost of building broadband infrastructure)

  • Individual can provides ground level recording from their location (challenges)

  • Fabric data

  • Upcoming FCC updates (API?)

  1. Sharing our results: Providing a reproducible way for others to explore FCC data at scale

Contacts

Website: https://ruralinnovation.us

LinkedIn | Twitter | Facebook | Instagram | YouTube

Check out our BEAD Climate Tool

https://broadband-risk.ruralinnovation.us/#about