Signal's investigations are built on publicly available, primary-source data — government filings, scientific databases, regulatory records, and open APIs. This registry documents the 40 most important datasets we use, why they matter, and how to access them. Every dataset listed here is either free to download or accessible via a free API tier. No paywalls, no proprietary sources.
No datasets match your search. Try broader terms or clear the filter.
$
Economic
10 sources
Fiscal records, subsidy databases, campaign finance filings, and financial crime data. Used in Signal's investigations into fossil fuel subsidies, lobbying ROI, cybercrime losses, and dark money flows.
#01 — Economic
IMF Fossil Fuel Subsidies Database
Explicit and implicit fossil fuel subsidies for 170 countries from 2010 to 2024. The definitive global dataset. Implicit subsidies (environmental costs not reflected in prices) dwarf explicit ones — total $7 trillion in 2022. Download includes per-country, per-fuel breakdowns in Excel format. Used by Signal in the G7 accountability investigation.
FBI's Internet Crime Complaint Center annual reports from 2015 to 2024. The only comprehensive US cybercrime loss dataset: $16.6 billion in losses in 2024 (+33% YoY), 859,000 complaints. Broken down by crime type, state, victim demographics. PDF + Excel. The foundational source for Signal's cybercrime investigation.
Federal Election Commission's complete PAC filing database — machine-readable, updated regularly, covering all registered PACs, super PACs, and hybrid PACs. The primary source for dark money and political spending investigations. API available. Free, no authentication required. Signal uses this cross-referenced with OpenSecrets to map lobbying ROI.
US Federal Election Spending 2000–2024
$14.4B (2024)
Annual lobbying expenditure by company and trade association broken out by bill and issue area, back to 1998. The most comprehensive aggregator of FEC and Senate lobbying disclosures. Oil Change International's analysis using this dataset showed fossil fuel industry political contributions vs. federal subsidy received — Signal's 37.8:1 ROI frame.
The most comprehensive database of state and local economic development subsidies — over 800,000 subsidy awards across the US, searchable by company, state, and incentive type. Includes the Virginia data centre tax exemption data used in Signal's Ashburn investigation. Shows that data centre incentives ballooned 1,051% since the program began.
Virginia's Joint Legislative Audit and Review Commission publishes annual state incentive spending — the source revealing that data centres consumed 80% of all Virginia economic incentive spending ($1.9 billion) in FY2025. Machine-readable tables in appendices. Free PDF download from jlarc.virginia.gov. The most authoritative accounting of Virginia's data centre deal.
Oil Change International — Federal Subsidies Analysis
Annual analysis of US federal subsidies to fossil fuel production, cross-referenced against Joint Committee on Taxation scorecards. The September 2025 report documented $34.8 billion/year in total federal fossil fuel support — including $4 billion/year added by the One Big Beautiful Bill Act. Free PDF. Source for Signal's lobbying ROI investigation.
Every US nonprofit's IRS Form 990 financial disclosure — searchable, downloadable. The 990 database is the only public window into 501(c)(4) "dark money" organisations that can spend unlimited amounts on political advertising without donor disclosure. ProPublica has made 3+ million 990s machine-readable. Essential for beneficial ownership chain investigations.
US Treasury's Office of Foreign Assets Control publishes the SDN list in machine-readable XML — updated in near real-time. Lists all individuals, entities, and vessels under US sanctions. Signal cross-references this with FinCEN beneficial ownership data and OpenCorporates to trace sanctioned entities within Western corporate structures. Free, no authentication.
The Corporate Transparency Act (2024) created a FinCEN database of beneficial ownership information for most US companies. Access is controlled (law enforcement primary) but journalists can request through FinCEN's law enforcement portal. Combined with OFAC SDN data and OpenCorporates, this is the most powerful corporate accountability dataset in US history.
Energy infrastructure filings, data centre regulatory records, grid capacity data, and technology deployment databases. Signal's primary sources for the data centre, AI infrastructure, and wind energy investigations.
#01 — Technology
JLARC Data Centers in Virginia (Dec 2024)
Virginia's Joint Legislative Audit and Review Commission 2024 report on the data centre industry — the most comprehensive public analysis of a state's data centre sector ever published. Includes power consumption data, grid stress analysis, water use, economic impact, and tax exemption accounting. Free PDF download. Signal's single most-used source for the Virginia investigation.
The International Energy Agency's analysis of data centre and AI energy consumption globally — projecting 536 TWh in 2025, doubling to over 1,000 TWh by 2030. Includes country-level data, capacity projections, and power source breakdowns. Free download with registration. The authoritative global benchmark for Signal's energy investigations.
Global Data Centre Energy 2010–2030
536 TWh (2025)
Dominion Energy's annual Integrated Resource Plan — regulatory filing with the Virginia SCC projecting electricity demand, generation mix, and capital investment through 2045 across multiple scenarios. The 2025 IRP explicitly acknowledges that meeting data centre demand while complying with Virginia's Clean Economy Act requires technologies "not yet proven viable." Free PDF.
National Renewable Energy Laboratory's complete database of all US wind turbines — location (lat/lon), capacity (MW), hub height, install year, project name. Over 70,000 turbines. Used by Parker & Winikoff (UW/USDA) to analyse land ownership patterns. Signal uses this cross-referenced with USDA land parcel data to map the inequality of wind lease income distribution.
IEA's October 2025 commentary and database on China's export controls on rare earth elements — includes market share by processing stage (China controls 94% of permanent magnet production), export volume data by country, and scenario analysis. CC BY 4.0 licence — freely reproducible with attribution. The authoritative source for Signal's rare earth supply chain investigation.
US Energy Information Administration's state-level electricity generation, consumption, and pricing data — updated monthly, going back decades. Virginia data shows data centres consumed approximately 25–30 TWh annually in 2025 — 35% more than Texas, which ranked second. Free API, no key required for basic data. Signal's source for the state-level energy comparison.
Top States by Electricity Consumption
~155 TWh VA (2024)
National Renewable Energy Laboratory's comprehensive analysis of wind turbine blade waste — projecting 2.2 million tons of US blade waste by 2050. Includes recycling technology readiness assessments, state policy comparisons, and decommissioning bond requirement analysis. Free PDF. The foundational source for Signal's "Blade Problem" investigation.
PJM Interconnection is the transmission organisation that powers the mid-Atlantic region including Virginia. Publishes load forecasts, capacity auction results, and interconnection queue data publicly. The 2024 PJM auction saw an 833% price increase for the 2025–2026 delivery year driven by data centre demand. Free public data portal.
Department of Energy's WINDExchange database tracks wind energy's economic impacts — lease payments, jobs, and tax revenues at state level. Documents $935 million paid to rural US landowners from wind leases in 2022 alone. Signal cross-references this with USDA land parcel data to investigate the inequality in wind lease income distribution between large and small landholders.
US Geological Survey's real-time and historical groundwater and streamflow data for every US county — free REST API, no authentication for basic queries. Signal uses this cross-referenced with data centre location data to show watershed stress levels in "Data Center Alley." Also used in the California drought forensics investigation (FM1000 fuel moisture is derived from USGS gauges).
Atmospheric science, drought monitoring, wildfire data, water stress indices, and natural catastrophe records. The backbone of Signal's climate investigations — from the LA fires forensics to the global water crisis.
#01 — Climate
NOAA NCEI Billion-Dollar Disasters
NOAA's National Centers for Environmental Information track all US weather and climate events with losses exceeding $1 billion — from 1980 to present. 2024 saw 27 such events totalling $182.7 billion. State-level summaries available. Free download and API. Signal used this to show California's 46 confirmed billion-dollar events since 1980 and the step-change from 2020–2024.
US Billion-Dollar Disasters 1980–2024
$182.7B (2024)
Weekly drought status maps for the entire US — from D0 (Abnormally Dry) to D4 (Exceptional Drought). Free API returning GeoJSON by county, weekly data from 2000 to present. Signal used this to document that Southern California was experiencing its driest start to a water year in 44 years by January 7, 2025 — days before the LA fires ignited.
World Resources Institute's basin-level water stress, seasonal variability, and groundwater depletion scores for every country — downloadable as GeoJSON for direct mapping. 17 nations already face "extremely high" water stress year-round. Signal uses this to map data centre locations against watershed stress and to identify geopolitical flashpoints in the global water crisis investigation.
NASA's Gravity Recovery and Climate Experiment Follow-On (GRACE-FO) satellite measures groundwater mass anomalies globally — the only independent source of aquifer depletion data that doesn't rely on corporate self-reporting. 2002–present. Signal uses this to independently verify what's happening to groundwater beneath hyperscale data centres built in drought-stressed regions.
NOAA's 1000-hour dead fuel moisture (FM1000) index — available back to the 1980s for climate stations across the US. By January 7, 2025, FM1000 in Southern California was at its 6th lowest on record for that date. Signal used this for the LA fires forensic reconstruction, showing the conditions were predictable from months of data.
First-of-its-kind public dataset: homeowner insurance non-renewals at county level for all 50 US states + DC, 2018–2023. Released under Senator Whitehouse's climate insurance investigation. Machine-readable CSV. Signal is building a national county-level choropleth map from this — no other outlet has visualised it at this geographic resolution.
US Homeowner Non-Renewals 2018–2023
+22% (2018–2023)
UN Food and Agriculture Organization's AQUASTAT tracks freshwater withdrawal by country and sector — agriculture, industry, municipality — going back to 1960 in some countries. The definitive source for long-term water consumption trajectories. Signal uses this with WRI Aqueduct to identify countries exceeding sustainable withdrawal rates, the precursor to water stress and conflict.
California Department of Insurance public filings on the FAIR Plan — California's insurer of last resort. The FAIR Plan tripled in size from 140,000 policies in 2018 to 610,000 by June 2025 as major insurers withdrew from wildfire-prone areas. County-level exposure data publicly filed. Signal cross-references this with Zillow property prices to map where climate change is visibly depressing property values.
The most comprehensive database of water-related conflicts, crises, and violence — updated annually since 1990. Covers 1,200+ documented incidents globally. Signal applies Signal's catastrophe modelling methodology to this: cross-referencing WRI Aqueduct stress scores with conflict incident timing to calculate average lead time between extreme water stress and armed conflict.
Aon's annual and semi-annual global natural catastrophe loss reports — the insurance industry's most authoritative public accounting of insured vs. total economic losses. H1 2025 insured losses reached $100 billion — 40% above H1 2024 and more than double the 21st-century average. Published via WEF. Free PDF. Signal's primary source for the climate insurance cascade investigation.
Global Insured Disaster Losses 2010–2024
$100B H1 2025
Global disease burden data and air pollution exposure estimates from the world's most comprehensive public health research programme. IHME's GBD tracks how pollution, disease and premature death are distributed across 204 countries — the gold standard for health-risk cross-referencing with climate, industrial and financial datasets.
#01 — Health
IHME GBD 2023 — Air Pollution Exposure & Disease Burden
The Global Burden of Disease Study 2023, coordinated by the Institute for Health Metrics and Evaluation (IHME), provides the most comprehensive public estimate of how air pollution kills. In 2023, air pollution contributed to 7.9 million deaths globally — the second-leading environmental risk factor after high blood pressure. The dataset covers PM2.5, ozone, nitrogen dioxide, and household air pollution for 204 countries from 1990–2023. Key outputs include deaths attributed to each pollutant, disability-adjusted life years (DALYs), and population-weighted exposure estimates. Gridded exposure data at ~10km resolution available for PM2.5 and NO2. Free download under IHME's non-commercial use agreement. Signal uses this to cross-reference industrial cluster locations (data centres, refineries, ports) against local PM2.5 burden — quantifying the human cost of infrastructure siting decisions.
The complete GBD Results interface covers all 371 diseases and injuries, 87 risk factors, and 22 causes of death across 204 countries and 811 subnational locations — 1990 to 2023. Outputs include incidence, prevalence, mortality, YLLs, YLDs, DALYs, and HALE (healthy life expectancy). The interactive VizHub tool allows custom downloads by cause, location, year, age, and sex without coding. For programmatic access, the GHDx (Global Health Data Exchange) provides bulk download of all source data. Signal applies GBD risk-factor burden data to financial datasets: mapping high-pollution counties against health insurance claim rates, hospital infrastructure, and corporate ESG disclosures — quantifying the financial liability that air quality creates for insurers and employers.
Cite: IHME. GBD Results. Seattle, WA: IHME, University of Washington, 2025. https://vizhub.healthdata.org/gbd-results/
🌡
CO₂ & Emissions
3 sources
The definitive open-access datasets on global carbon emissions, fossil fuel consumption, and energy transition progress. From the Global Carbon Project's annual carbon budget to the Energy Institute's Statistical Review — the primary sources used by the IPCC, central banks, and climate litigators.
#01 — Emissions
Global Carbon Budget 2025 — GCP Fossil & Land-Use CO₂
The Global Carbon Project's annual carbon budget is the most widely cited accounting of all anthropogenic CO₂ emissions — updated every November at COP. The 2025 release (GCB2025, Friedlingstein et al.) covers global fossil CO₂ emissions (EFOS), land-use change emissions (ELUC), the atmospheric growth rate, and ocean and land carbon sinks through 2024. Total anthropogenic emissions reached 41.6 GtCO₂ in 2024 — a record. The remaining 1.5°C carbon budget from 2025 is approximately 235 GtCO₂, equivalent to around 6 years at current emissions. Country-level time series are available as free CSV/XLSX from ICOS Data Portal and Zenodo. The Our World in Data processing pipeline (owid/co2-data on GitHub) provides cleaned, country-standardised CSVs covering all nations from 1750–2024 with consistent methodology. Signal uses this as the baseline for all climate-financial risk quantification — the carbon constraint is the foundational input into physical and transition risk modelling.
Global Fossil CO₂ Emissions 1960–2024
37.4 GtCO₂ (2024)
Cite: Friedlingstein et al. (2025). Global Carbon Budget 2025. Earth System Science Data. DOI: 10.5194/essd-2025 (forthcoming). Data: globalcarbonbudget.org
#02 — Emissions
Energy Institute Statistical Review of World Energy 2025
The Energy Institute's Statistical Review — formerly the BP Statistical Review, now independently published — is the global energy industry's annual reference dataset. The 2025 edition (74th) covers production, consumption, trade, prices, and emissions for oil, gas, coal, nuclear, hydro, wind, solar, and other renewables across ~80 countries from 1965 to 2024. Key 2024 findings: global energy demand rose 2%, fossil fuels remained at 86% of the primary energy mix, CO₂ from energy hit 40.8 GtCO₂e (record for the fourth consecutive year), wind and solar grew 16% but didn't offset rising fossil demand. The full dataset is downloadable free as Excel (XLSX) and is the primary source behind Our World in Data's energy charts. The new methodology reports renewable energy at actual calorific content (not input-equivalent) — a significant methodological change that reduces headline renewables share. Signal uses this for the fossil fuel subsidy and energy transition investigations.
Fossil Fuel Share of Primary Energy
86% (2024)
Free XLSX download1965–2024~80 countriesAnnual · June release
Key figures (EI Statistical Review 2025, data year 2024)
86%fossil fuel share of global primary energy mix (2024)
40.8GtCO₂e energy-related emissions — 4th consecutive record
+16%wind & solar generation growth in 2024
+2%total global energy demand growth in 2024
109 Gtcumulative emissions avoided by renewables + nuclear since 2010
Cite: Energy Institute (2025). Statistical Review of World Energy, 74th edition. London: Energy Institute. energyinst.org/statistical-review
#03 — Emissions
Our World in Data — CO₂ & GHG Emissions Dataset
Our World in Data's open CO₂ and Greenhouse Gas Emissions dataset aggregates and standardises the Global Carbon Budget, the Energy Institute Statistical Review, and Jones et al. national GHG contributions into a single machine-readable file — the most analysis-ready form of these sources. Covers CO₂, methane, nitrous oxide, and total GHGs; per capita, per GDP, cumulative, and consumption-based (trade-adjusted) variants; by country and region from 1750 to 2024. All code is open-source on GitHub (owid/co2-data). This is the dataset Signal's pipeline uses for rapid country-level emissions cross-referencing — plugged directly into the visualise tool. Free CC-BY licence for all OWID-processed data.
Cite: Hannah Ritchie, Pablo Rosado, Max Roser (2023). CO₂ and Greenhouse Gas Emissions. OurWorldInData.org. Data: Global Carbon Project; Energy Institute; Jones et al. (2024).
⚔
Defence & Geopolitics
2 sources
Military expenditure and arms transfer data from the Stockholm International Peace Research Institute — the world's foremost independent authority on global security spending. The definitive source for tracing the financial flows of geopolitical tension: who is spending, how fast, on what, and where the money comes from.
#01 — Defence
SIPRI Military Expenditure Database 2025 (1949–2024)
The Stockholm International Peace Research Institute's Military Expenditure Database is the global standard for defence spending data — covering 173 countries from 1949 to 2024, updated annually each April. The 2025 release documents a historic inflection: global military spending reached $2,718 billion in 2024 — a 9.4% real-terms increase, the steepest single-year rise since at least 1988. Spending has increased every year for a full decade, up 37% since 2015. NATO's 32 members spent $1,506 billion in 2024 — 55% of global total.
World Military Spending 1988–2024
$2,718B (2024)
Free Excel download173 countries1949–2024Annual · April release
Cite: SIPRI Military Expenditure Database, April 2025. DOI: https://doi.org/10.55163/CQGC9685
#02 — Defence
SIPRI Arms Transfers Database (TIV, 1950–2024)
SIPRI's Arms Transfers Database tracks the flow of major conventional weapons between countries using Trend Indicator Values (TIVs) — a standardised measure of military capability transferred, not financial value. This distinction is critical: TIVs allow comparison of weapons transfers regardless of actual contract prices (which are often classified). The database covers over 70 years of arms flows across 190+ exporters and importers. Key 2024 trends: global arms transfers rose for the fifth consecutive year; Europe's share of global imports reached its highest level since the Cold War (driven by Ukraine, Poland, Germany); the US retained its position as the world's largest arms exporter with a 43% share. India remained the largest single importer. Free download as CSV. Signal uses this cross-referenced with SIPRI Milex and SEC EDGAR defence contractor 10-K filings to connect arms transfer flows to revenue at Lockheed Martin, RTX, Northrop Grumman and others — the financial anatomy of geopolitical conflict.
Global Arms Transfers (TIV) 2005–2024
+↑ 5th consecutive yr
Cite: SIPRI Arms Transfers Database. Stockholm International Peace Research Institute. sipri.org/databases/armstransfers
⬡
Other
10 sources
Geopolitical risk, corporate registry, land rights, health, and social infrastructure data. The miscellaneous-but-critical sources that power Signal's cross-cutting investigations into financial crime, sanctions, and inequality.
#01 — Other
USDA NASS Census of Agriculture — Land Parcel Data
US Department of Agriculture's National Agricultural Statistics Service Census of Agriculture provides land parcel size distribution by county — the foundational dataset for Signal's wind farm inequality investigation. Shows that 53% of Wisconsin rural tracts are under 100 acres (too small for wind developers to approach) vs. 35% in Iowa, explaining the state's lower wind development rate.
The largest open database of companies in the world — 200+ million corporate entities across 130+ jurisdictions, with API access. Free tier available. Signal uses this cross-referenced with OFAC SDN data to trace sanctioned entities through shell company chains in Delaware, Wyoming, and the Cayman Islands. Essential for beneficial ownership and sanctions evasion investigations.
The International Consortium of Investigative Journalists' searchable database of entities and addresses from the Panama Papers, Paradise Papers, Pandora Papers, and other major financial leaks. Over 800,000 offshore entities. Free search and bulk download. Signal uses this for sanctions evasion investigations — identifying known offshore structures that may link to OFAC-listed entities.
Comprehensive tracking of global offshore wind project status — including cancellations. Documents the jump from 2 cancelled projects in 2021 to 50+ in 2024. Tracks developer-initiated vs. government-driven cancellations. US accounts for 30% of total cancelled capacity globally. Free web access for published analysis; Signal's primary source for the "Stop-Work Files" investigation.
Institute for Energy Economics and Financial Analysis publishes rigorous financial analysis of energy projects. January 2026 report: offshore wind stop-work orders could cost consumers billions and delay grid-critical capacity. South Fork Wind capacity factor averaged 53% for Nov 2024–Apr 2025, outperforming the average US coal plant (46%). Free PDF reports. Signal's source for consumer cost impact modelling.
Zillow publishes property price indices at zip code level — free download. Signal cross-references this with California DOI FAIR Plan concentration data to identify zip codes where the insurance crisis is already depressing property values. The data that allows Signal to quantify Jerome Powell's warning: "there will be regions where you can't get a mortgage."
Bureau of Ocean Energy Management's public database of all offshore wind leases, permits, and construction statuses on the US Outer Continental Shelf. The source for construction progress data on Revolution Wind (87% complete at stop-work), Empire Wind, Sunrise Wind, and Coastal Virginia Offshore Wind. Updated monthly. Free GIS data download including lease area shapefiles.
US Geological Survey publishes annual Mineral Commodity Summaries with global production and reserve data for all minerals including rare earth elements — the authoritative non-industry source for supply data. Shows China controls over 60% of global REE mining despite holding only ~35% of known reserves — the processing dominance story. Free PDF and Excel.
EPA's Enforcement and Compliance History Online tracks environmental violations by facility — searchable by company, location, and violation type. National dataset. Signal plans to use this cross-referenced with data centre locations and tax abatement records to find the Oregon pattern (Amazon illegal water dumping / fine waivers) replicated nationally.
Environmental and Energy Study Institute analysis of Northern Virginia data centre water consumption, drawing on ICPRB (Interstate Commission on the Potomac River Basin) water utility data. Documents that data centres in Northern Virginia consumed close to 2 billion gallons of water in 2023 — a 63% increase from 2019. Signal's primary source for the Ashburn water reckoning.