Back to Delos

TECHNICAL SPECIFICATION · v2.1

The Signal Score Methodology

A reproducible, multi-source scoring system for detecting technology infrastructure gaps in underserved communities globally.

Core Formula

Signal Score = 0.30 × MDG + 0.25 × EES + 0.20 × IAS + 0.15 × SVS + 0.10 × TDS

All sub-scores are normalized 0-100. Satellite flags (SAR, GRACE-FO, NDVI, LST) can independently elevate a brief's priority regardless of composite score.

01

Sub-Score Definitions

Each sub-score is computed independently from publicly available data sources and normalized to a 0-100 scale before weighting.

MDGMonitoring Density Gap30% weight

Measures the gap between a community's environmental risk level and the density of physical monitoring infrastructure (gauges, sensors, stations) actively measuring it.

MDG = Risk_Composite × (1 − Monitoring_Composite / 100)

Risk is drawn from flood/fire/drought exposure. Monitoring is computed from USGS gauge density within 50km radius.

Data Sources

DatasetUpdate
USGS National Water Info SystemReal-time
FEMA National Flood Hazard LayerQuarterly
NASA FIRMS Fire InformationDaily
US Drought MonitorWeekly
Low

0-30: Well-monitored for its risk level

Mid

30-60: Moderate monitoring gap

High

60-100: High-risk area with minimal sensor infrastructure

EESEnvironmental Exposure Score25% weight

Quantifies cumulative environmental hazard burden across six dimensions: flood risk, wildfire exposure, drought severity, storm event history, air quality, and toxic contamination.

EES = weighted_average(Flood, Fire, Drought, Storm, Air, Toxic) × 100

Flood risk (FEMA) carries highest weight at 0.30; air quality (EPA AQI) and toxics (EPA EJScreen) contribute equally at 0.15 each.

Data Sources

DatasetUpdate
EPA EJScreenAnnual
NOAA Storm Events DatabaseMonthly
FEMA National Flood Hazard LayerQuarterly
NASA FIRMSDaily
Low

0-30: Low cumulative hazard burden

Mid

30-60: Moderate environmental exposure

High

60-100: High cumulative hazard - frequent multi-hazard overlap

IASInfrastructure Access Score20% weight

Measures gaps in the physical and digital infrastructure that communities need to receive and operate technology solutions: broadband coverage, healthcare facility access, and grid reliability.

IAS = (1 − Broadband_penetration) × 0.40 + Healthcare_gap × 0.35 + Grid_unreliability × 0.25

Broadband from FCC Form 477; healthcare gap from CMS rural health data; grid reliability from EIA outage records.

Data Sources

DatasetUpdate
FCC Broadband Data CollectionSemi-annual
HUD Housing Needs AssessmentAnnual
USDA Food Access Research AtlasAnnual
US Census TIGER RoadsAnnual
Low

0-30: Good infrastructure access

Mid

30-60: Partial access gaps

High

60-100: Severe infrastructure deficit - deployment is hard

SVSSocioeconomic Vulnerability Score15% weight

Captures the human vulnerability dimension: income poverty, health burden, housing cost, unemployment, and linguistic isolation - factors that compound when technology gaps exist.

SVS = CDC_SVI_composite × 0.50 + ACS_poverty_rate × 0.30 + CDC_PLACES_health × 0.20

CDC SVI is the primary composite; ACS income and PLACES chronic disease burden are secondary weights.

Data Sources

DatasetUpdate
CDC Social Vulnerability IndexBiennial
CDC PLACESAnnual
US Census ACS 5-YearAnnual
Low

0-30: Lower vulnerability - adoption support needs are modest

Mid

30-60: Moderate vulnerability

High

60-100: High vulnerability - solutions need local partner organizations

TDSTechnology Deployment Score10% weight

A measure of how historically underinvested this community has been in technology solutions - estimated from SBIR award history, broadband investment patterns, and IoT sensor density.

TDS = (1 − SBIR_award_density) × 0.60 + Low_sensor_density × 0.40

SBIR awards geocoded from SBIR.gov; sensor density estimated from USGS + EPA monitoring station locations.

Data Sources

DatasetUpdate
SBIR.govMonthly
NSF Award SearchMonthly
Low

0-30: History of tech investment

Mid

30-60: Some prior investment

High

60-100: Chronically underinvested - greenfield opportunity

02

Satellite Intelligence Layer

Four independent satellite data streams run in parallel to the ground-based scoring. A positive satellite flag can independently trigger a brief regardless of composite Signal Score.

🛰

SAR Coherence (Sentinel-1)

ESA Sentinel-1 C-band SAR · Every 6-12 days

Coherence drop >0.15 between consecutive passes indicates significant ground change - structural collapse, rapid construction, or severe vegetation loss.

Alert threshold

Coherence < 0.35 in built-up areas triggers SAR Alert flag

💧

Groundwater Anomaly (GRACE-FO)

NASA/DLR GRACE-FO satellite pair · Monthly

Terrestrial water storage anomaly from gravity measurements. Persistent depletion below −5 cm/year over 3+ years indicates aquifer stress.

Alert threshold

Anomaly < −5cm AND trend declining > 3 years triggers Groundwater Alert

🌿

Vegetation Stress (MODIS NDVI)

NASA Terra MODIS · 16-day composite

Normalized difference vegetation index tracks agricultural health and vegetation stress. Sustained anomaly below 1.5σ from baseline indicates drought stress or land degradation.

Alert threshold

NDVI anomaly < −1.5σ from 5-year baseline over 60 days

🌡

Urban Heat Island (Landsat)

USGS/NASA Landsat 8/9 TIRS · Every 16 days

Land surface temperature offset between urban core and peri-urban buffer zone. Persistent delta >4°C indicates elevated heat risk for vulnerable populations.

Alert threshold

LST urban-buffer delta > 4°C in summer months

03

Brief Generation Triggers

A brief is generated when one or more trigger conditions are met. Each brief includes a trigger attribution so readers know exactly what caused it.

TriggerSource
Score ThresholdScoring Engine
SAR AlertGoogle Earth Engine
Groundwater AnomalyNASA PO.DAAC
NDVI StressNASA MODIS
Heat IslandLandsat TIRS
Dark ZoneSentinel-1 + Census
News EventGDELT 2.0
Community ReportCommunity Portal
04

Accuracy and Validation

Multi-source corroboration is required before a brief is published. A single anomalous data point cannot trigger publication.

Multi-source corroboration

A minimum of 2 independent data sources must agree before a brief is published. Satellite alerts require ground-data corroboration.

Required

Data freshness scoring

Every data point carries a freshness timestamp. Briefs display per-field data currency so readers can judge reliability.

Automated

Community validation

Residents and local organizations can flag, confirm, or correct briefs. Confirmed briefs carry a distinct badge and rank higher in search.

Human-in-loop
05

API Reference

The Signal Score API is publicly accessible. All endpoints return JSON.

GET/score/tract/{fips}
POST/score/point
GET/score/ranking
GET/briefs
POST/briefs/search
GET/briefs/{id}
POST/community/reports
POST/founders/claim

Citing Delos

Caelus Research (2026). Delos: A Satellite-Powered Infrastructure Gap Discovery Platform. signal.caelus.org. Methodology v2.1.

All Signal Scores and briefs are freely reproducible from publicly available data. If your research relies on Delos data, please cite appropriately and consider contributing community validations.