The Connect Score

What it measures

The Connect Score is a parish-level composite of three components, each answering a distinct question about digital equity:

• Availability, can residents get modern broadband? Pulled from the FCC Broadband Data Collection: the share of locations with at least one provider offering 100/20 Mbps service.
• Affordability, can they afford it? A combined index of Lifeline subscriber take-rate and structural cost-burden, computed from American Community Survey income data.
• Adoption, do they use it? American Community Survey table B28002 broadband subscription rate among households.

How the components combine

The published v1.0.0 weights:

Adoption gets the most weight because it is the closest single measure of the score’s construct, residents being equitably connected. Affordability is weighted heavily because it is the biggest remaining barrier among households where broadband is technically available. Availability earns the smallest share because, in 2026 Louisiana, roughly 95% of locations are already served at 100/20 Mbps, so the marginal point of availability buys less digital equity than the marginal point of affordability or adoption.

Bands and colors

Scores are grouped into five bands using Jenks natural breaks, an algorithm that finds the most distinct clusters in the actual distribution of Louisiana parishes. The colors come from ColorBrewer’s RdYlGn 5-class colorblind-safe palette, chosen so the score stays distinguishable for the most common forms of color vision difference.

How we validated it

Every published version of the Connect Score ships with three layers of independent validation: internal consistency, convergent validity against poverty, education, and rurality, and criterion validity against Louisiana’s own BEAD investment priorities.

The headline finding for v1.0.0is that the Connect Score predicts the same parishes Louisiana’s department prioritized for $1.4 billion in BEAD investment, with strong negative correlation. Read the full reports:

Data sources

The v1.0.0 score uses these data vintages:

How the score changes over time

The composite Connect Score is a snapshot, labeled by version. We do not claim year-over-year comparability for the composite itself. If you want to track Louisiana’s progress over time, the underlying raw inputs, percent served and percent subscribed, are the longitudinal record. When the methodology changes, prior scores stay published as-is under their original version, and the new version’s release notes document what changed.

v1.0.0 exception: this version received same-day display-layer corrections on 2026-04-27. Band thresholds were rounded to whole numbers so two parishes displaying the same score cannot land in different bands, and one mid-yellow swatch was swapped for visibility. Four parishes re-banded as a result. Underlying scores did not change. Future post-launch methodology changes will bump to a new version rather than mutate in place.

Limitations and what’s next

• Place-level (city) scores are not published in v1.0.0 because household-survey data has wide margins of error at small geographies. Place pages show the containing parish’s score with clear labeling.
• v1.0.0estimates Lifeline-eligible households as a flat 20% of each parish’s total households — a statewide proxy, not parish-specific poverty data. This compresses the Lifeline take-rate signal in the highest-need parishes (where the true eligible share is higher than 20%) and inflates it in the lowest-need parishes. The 40/60 structural/Lifeline internal-split sensitivity (Spearman 0.915, 21 of 64 parishes shifting bands) is the closest stress test on this assumption. Future versions will source eligibility directly from ACS poverty data.
• Quality of available service (technology mix) and provider competition are preserved in the per-parish data but do not contribute to the v1.0.0 score. A future version may bring them back if the construct evolves.