Markets, the national debt, and the same year-to-date signals that run in the geopolitical report — straight from Treasury, FRED and the exchanges. No spin, just the arithmetic.
Live index levels versus the prior close. Refreshes every two minutes.
Total public debt outstanding, extrapolated live from the last Treasury settlement at the trailing-year accrual rate.
What a dollar is worth abroad — and what it still buys at home. Index quotes are live; the cross rates are the European Central Bank’s daily reference fixing.
The other half of the story. Every figure is the CPI-U ratio — the same index the government uses to index its own obligations.
The 10-year breakeven — the inflation rate that makes a regular Treasury and an inflation-protected one pay the same. No survey, no BLS methodology debate: this is what people betting real money think inflation will average over the next decade.
Arthur Okun’s kitchen-table gauge: the unemployment rate plus year-over-year inflation, one number, monthly since 1948. No modeling, no weighting debates — just the two ways a government most reliably makes ordinary life worse, added together. Refreshes every six hours.
The charts that ship with the daily geopolitical report — crime, rates & credit, labor, housing, commodities and the dollar.
Canada publishes a body count for people who died waiting for care. The United States does not — not because our waits are shorter, but because we never built the list. We don’t ration by queue; we ration by price and by door, and nobody is required to count what that costs. So this is the closest thing the federal government actually publishes: the people who sat in an emergency-room waiting area and walked out before anyone looked at them, how long the ones who stayed sat there, and how many Americans never made it to a waiting room at all. Every figure below is CMS’s or CDC’s own. Refreshes twice a day.
Disease surveillance straight from CDC’s open-data feeds — the sewers get sampled whether anyone likes the answer or not. Wastewater viral activity, measles case counts, and weekly deaths. Refreshes every six hours.
The same discipline, world-scope — WHO outbreak bulletins as they publish, influenza positivity from every FluNet reporting lab on the planet, and live GDACS disaster alerts. Refreshes every six hours.
Provisional drug overdose deaths — rolling 12-month counts straight from CDC vital statistics, national and by state. Half a century and a trillion dollars into the drug war, the chart below is the scoreboard. The recent drop is the steepest ever recorded; decide for yourself what moved it. Refreshes every six hours.
Twenty-one causes of death, age-adjusted rates per 100,000, straight from CDC vital statistics — the whole American mortality ledger on one page, updated quarterly. GVA counts shooting incidents; this counts bodies per capita, the official cross-check. Refreshes every six hours.
Deaths from pregnancy per 100,000 live births, rolling 12-month windows — the number that separates the United States from every peer nation, published monthly by CDC’s own statisticians. The race gap below is not commentary; it is the agency’s own rows. Refreshes every six hours.
The measles pane above shows the virus arriving in the sewers. This shows the door it walks through: MMR coverage among kindergartners, state by state, against the 95% community-immunity line CDC itself cites — and the exemption paperwork holding it open. Refreshes every six hours.
The medical-industrial complex in the government’s own numbers — FDA device-death reports, FDA drug-death reports, the hospitals CMS grades worse than the national rate, and the latest malpractice opinions from the federal courts. These are only the cases anybody bothered to report. Refreshes every six hours.
Every report a manufacturer, hospital, or doctor filed with the FDA’s MAUDE database this year. “Malfunction” means the device failed; “injury” and “death” mean it failed into a person.
Which devices show up most often in this year’s death reports.
Which drugs are named most often in this year’s fatal adverse-event reports. Named in a report ≠ proven cause — but the pile is the pile.
CMS grades every Medicare-certified hospital against the national rate on deaths and complications. These are the measures where hospitals are currently graded worse — the government’s own scorecard, not ours.
The most recent published court opinions mentioning medical malpractice, via CourtListener.
Every FDA food-recall enforcement report since 2012, by year — how many recalls, how many were Class I (reasonable probability of serious harm or death), and how much food by weight. Straight off the openFDA enforcement API, no intermediary. Refreshes every six hours.
Every foodborne outbreak reported to the CDC since 1998 — how many got sick, how many landed in a hospital, how many died, and which pathogens did it. Straight off the National Outbreak Reporting System, no intermediary. CDC estimates ~48 million foodborne illnesses a year; these charts count only the ones traced to a reported outbreak, so every line is a floor. The reporting system runs two to three years behind the calendar — the newest complete year here is what CDC has actually finished counting.
What it costs to carry the national debt, fiscal-year-to-date, out of the Treasury’s own books. The headline is interest paid to actual bondholders — interest the Treasury credits to its own trust funds is broken out separately, because that is a bookkeeping entry, not a cheque. Every year is measured at the same fiscal month, so a partial year is never set against a finished one. Refreshes every six hours.
What lenders charge the Treasury at every maturity, straight from the Treasury’s own daily par yield curve. The curve is the price tag on the Interest Meter above — every dollar of debt rolls over at these rates. When short-term paper pays more than long-term paper the curve is “inverted”, and an inversion has preceded every US recession since the 1970s. Refreshes every six hours.
Every marketable Treasury security, CUSIP by CUSIP, sorted by the year it dies. Trillions were borrowed in the near-zero era at coupons under 3% — and every dollar of it gets reborn at today’s rates on a date already printed on the bond. Coupon averages cover nominal notes and bonds only; TIPS and floating-rate paper are counted in the totals but excluded from the averages. Refreshes with the Monthly Statement of the Public Debt.
The Federal Reserve’s balance sheet, the money supply it feeds, and the overflow parking lot. Every line here is the Fed’s own data. The balance sheet chart is the whole QE story in one stroke — 2008, 2020, and the slow grind back down. The M2 growth chart shows the fastest expansion of the money supply in American history, followed by the first outright contraction since the 1930s. Refreshes every six hours.
The charts above show the printing. This shows what it bought you — the same sandwich, priced twice a year since 2000 by The Economist. One burger, no substitution effects, no hedonic adjustments, no basket rebalancing. Just the price. Refreshes daily.
The tabs above this one are the government’s balance sheet. This one is yours — what families save, what they owe on the cards, what share of card debt is going unpaid, and what it costs to buy the roof. Refreshes every six hours.
Real median weekly earnings — the middle worker’s paycheck with inflation stripped out — against real GDP per person, both indexed to 1979. The gap between the two lines is the whole argument about where the growth went. Refreshes every six hours.
Where the jobs actually went — cumulative payroll change since January 2022, sector by sector, straight from the BLS establishment survey. Nine million jobs were added in total, but the ledger is wildly lopsided: nearly half the net gain is health care alone, and two columns run red. Government-adjacent sectors — health care, private education, state & local government — account for the majority of everything added since 2023. Recent months are preliminary and get revised twice. Refreshes daily.
Every other tab on this page measures what government spends. This one measures what it bought. Grade 8 national public-school scores from the NAEP Data Service — the federal government’s own report card, and the only test given the same way for thirty years. Refreshes twice a day.
The white–Black score gap is the number every school board in America is judged on. It has narrowed. Read the second chart before you decide that is good news — the gap closed because every group fell, and the group that was ahead fell faster.
Real per-pupil spending against test scores, both indexed to their first common year so they share one axis. Money is inflation-adjusted, so this is not a story about the dollar getting weaker. Spending runs through the Census Bureau's latest school-finance survey, the same year as the newest scores.
Fifty-one school systems, one test. Each dot is a state; the columns below are sortable — click a header. If money bought scores, the dots would climb from left to right. They don’t. The District of Columbia has the highest GDP per capita in the country and one of the lowest math scores in it.
College tuition against headline inflation and the hourly wage of a production worker, all indexed to 1978. Wages roughly kept pace with prices. Tuition did not — and the gap between those lines is now carried as household debt.
Both subjects peaked in 2013 and have fallen every assessment since. The slide begins six years before a single school closed for COVID — the pandemic accelerated a decline that was already underway, it did not start it.
The same test, sorted by how far the child’s parents got in school. The spread between a child of college graduates and a child of high-school dropouts is wider than the racial gap in the first chart — and no school-funding formula touches it.
How much of all personal income in America arrives as a government transfer — Social Security, Medicare, Medicaid, unemployment, veterans’ benefits, SNAP. Numerator and denominator both come from the same BEA vintage, so the ratio is honest. State figures are place-of-residence. Refreshes daily.
The Code of Federal Regulations, counted. How many sections carry obligation language — shall, must, may not, prohibited, required — which titles carry the most of it, and how much of it gets rewritten each year. These are counts of sections containing a term, not raw word counts; a section saying shall nine times counts once. Refreshes twice daily.
Federal disaster declarations as they issue, every live National Weather Service alert, and significant earthquakes — FEMA, NWS and USGS open feeds, no filter. Refreshes every three hours.
Every CBP encounter nationwide — the monthly ledger the arguments are supposed to be about. Straight from CBP’s published encounter data, updated monthly.
The same CDC death certificates, split by what killed. Fentanyl is the line that built the mountain and the line falling fastest. Heroin is a ghost of what it was. Meth and cocaine are dropping but slower. Categories overlap — a death involving both fentanyl and cocaine is counted in both rows — so the total is not the sum of the parts. Every point is a rolling 12-month count ending that month. Refreshes every six hours.
National offense rates from the FBI’s Crime Data Explorer — violent crime, property crime and homicide per 100,000 people, monthly. The federal numbers lag a year or more; that lag is part of the story.
Live facility-level counts from the Bureau of Prisons. This is the federal system only — it excludes state prisons, local jails and ICE detention, which together hold several times more people. Facility type is parsed from the name; complex rollups are excluded so nobody gets counted twice. Refreshes every three hours.
The Crime Watch pane above waits a year or more for the FBI. This one doesn’t: provisional homicide counts straight off American death certificates, monthly, running about four months behind the calendar — the freshest body count any federal agency publishes. And one number the tiles refuse to soften: Black Americans are roughly 14 percent of the population and nearly half the homicide victims in the FBI’s ledger. That disproportion — not the intra-racial percentages below, which everyone shares — is the actual asymmetry in American homicide. Refreshes every six hours.
“Black-on-black crime” is a phrase built on showing you one row of a table. Here is the whole table. Homicide is intra-racial for everyone, because violence is proximate — people are killed by the people near them: partners, rivals, neighbors, acquaintances. Ninety percent of Black victims are killed by Black offenders and eighty percent of white victims are killed by white offenders, and nobody has ever coined a phrase for the second number. Every race’s row, side by side, from the FBI’s own matrix — read them together or don’t read them at all.
For seventeen years the federal government asked every police agency in America how many gangs were operating on its streets, and published the answer. The last answer, 2012: more than thirty thousand gangs, eight hundred fifty thousand members, roughly one homicide in eight. Then the survey was defunded, the FBI’s gang threat assessments went dark, and the counting simply stopped — every “gang violence” number quoted today is a re-heated statistic from a country that no longer exists. This ledger ends where the funding did; the flatline after 2012 is the finding.
The Washington Post’s school-shootings database — every act of gunfire at a primary or secondary school during school hours, Columbine to now. The definition is deliberately narrow: no colleges, no after-hours shootings, no private suicides, no stray rounds that hit nothing. What remains is the thing people actually mean by “school shooting” — and most of them are not massacres. The median shooter is sixteen; the modal incident is a targeted dispute that kills nobody. The children were there for all of it: nearly half a million students were on campus while the shooting happened. This ledger lives apart from Domestic Extremism because the data says it belongs apart — overwhelmingly no political motive, mostly disputes, domestic violence and accidents that happened where children learn.
Two lines that refuse to move together. Incidents climb across the 2010s and spike after the pandemic reopening — but deaths are a handful of vertical walls: Sandy Hook, Parkland, Uvalde. Most years the death toll is in single digits across the entire country; the years everyone remembers are the years one indiscriminate shooter found an unlocked door.
The Post codes every incident by what the shooter was doing. Targeted disputes outnumber indiscriminate attacks four to one — but the indiscriminate column owns most of the graves. The armed-officer split is shown only where the Post recorded it, and it resolves nothing by itself: an officer was on campus at Parkland.
Who fires a gun at a school. The Post records the shooter’s age, gender, race and tie to the school where investigators established them, and this page states the known counts out loud — because for race the unknowns are nearly two-thirds of the ledger, and pretending otherwise would be the exact inflation this site exists to refuse. What the knowns say: overwhelmingly male, median age sixteen, most often a current student — and the gun came from home more than three times as often as from a lawful purchase by the shooter. One count nobody expects: in twenty-six incidents the person firing was a police or security officer on campus — mostly targeted fire in a confrontation, five accidental discharges, twelve dead. The database publishes no names; that is the Post’s editorial choice, and most of these shooters are juveniles whose records are sealed. The deadliest are different — publicly identified in court and investigative record, and named in the ledger above.
Forty-seven states have had gunfire at a school since Columbine. Raw counts, not per-capita — big states lead partly because they are big, same rule as the extremism maps. And the two columns refuse to agree: California owns the most incidents with twelve dead in twenty-six years; Texas is second on incidents and first on graves, because Uvalde and Santa Fe are single mornings that outweigh decades of disputes.
Las Vegas. Pulse. El Paso. Lewiston. Until now, none of them had a home on this site — the school ledger stops at the campus gate, and the terrorism databases require an ideology, which the deadliest shooting in American history never had. This is Mother Jones’ database of indiscriminate public rampages, 1982 to now: the narrow definition — a lone attack in a public place, gang hits and family murders excluded by design. Three or four dead is the entry fee depending on the year, and that threshold change is charted below, not hidden. What the narrow count shows: about seven a year, two-thirds by legally purchased guns, and the worst of them owning more graves than whole decades of everything else combined.
Two lines, and the second one is why this section exists. Incident counts crawl along at a handful a year for three decades, tick up after 2012 — and the death line spikes wherever one man found a crowd: Killeen 1991, Vegas 2017, the year sixty died at a concert and the chart never looks the same again. The 2013 mark is the definition change: the federal threshold dropped from four dead to three, so the right half of the chart counts slightly smaller tragedies than the left half. Most databases hide that seam. We draw it.
Workplaces and schools are the named categories; the biggest bucket is “Other” — concerts, nightclubs, supermarkets, a bowling alley — because a massacre goes wherever a crowd is. The deadliest ledger is named in full: these perpetrators were publicly identified in court and investigative record, and this site does not launder them into anonymity.
A different animal from the school ledger above, and the contrast is the finding. The school shooter is a sixteen-year-old current student with his father’s gun. The rampage shooter is a thirty-three-year-old man — 96% male, median age twice the school ledger’s — who bought his guns legally: nearly two-thirds of these arsenals passed every check on the books. And in half the cases someone saw warning signs before the first shot. Race roughly tracks the country’s demographics, with white shooters the majority and every group represented — the profile that predicts a rampage is not an ethnicity, it is a man with a grievance and a clean background check.
In 2023, Mother Jones counted twelve mass shootings. The Gun Violence Archive counted six hundred sixty. Both numbers are honest; both are on cable news; they are counting different things with the same words. Mother Jones counts indiscriminate public rampages — the Vegas kind. GVA counts every incident where four or more people are shot, killed or wounded, any venue, any motive — which sweeps in gang crossfire, drive-bys and family murders. Pick your definition and your number multiplies by fifty-five. This chart draws both lines on one canvas, and the flatline at the bottom is not an error — it is the entire Mother Jones series, dwarfed to the floor by GVA’s mountain. Whoever quotes you a mass-shooting count without naming their definition is running a con.
The number every parent needs and no cable segment gives: school is statistically the rarest place a child gets shot. The Post’s entire twenty-six-year school ledger — every campus, Columbine to now — holds 216 dead. The Gun Violence Archive counts more minors shot dead than that every seven weeks: over seventeen hundred children and teens in 2023 alone, overwhelmingly in homes and streets, not hallways. And the CDC’s own instrument says the quiet part: firearms are the leading cause of death for American children. The fear points at the schoolhouse; the dying happens everywhere else.
Every incident in the ADL’s H.E.A.T. map — hate, extremism, antisemitism and terrorism, 69,000-plus rows pulled nightly from the map’s own data feed and filed here under an honest name. About a fifth of what this database counts is crime. The rest is harassment that mostly cannot be charged, and propaganda drops and rallies that are protected speech outright. This page charts both without pretending either is the other. And because the ADL tracks no domestic terrorism at all — its feed is murders, plots and hate incidents, with no attack-level terrorism category — this tab folds in the instrument that does: the Global Terrorism Database’s U.S. ledger, 1970–2020, merged into the same charts, tiles and ideology list — one count, two instruments, nothing tallied twice.
Murders, terror plots and police shootouts back to 2002; assaults since systematic tracking began in 2016. Real crimes, real bodies — the part of this ledger nobody disputes belongs in it. The ideology list combines both instruments in the same buckets: the ADL’s own tags on its lethal-violence rows, plus every GTD U.S. perpetrator group mapped by hand onto those same labels — FALN and the Weather Underground under Left Wing, the Klan and the Aryan groups under White Supremacist, the militia and sovereign-citizen cluster under Anti-Government, anti-abortion violence under Right Wing (Other), jihadist attacks under Islamist. The Cuban-exile and JDL bombing campaigns of the 1970s file under Other, because left/right is the wrong axis for them.
The violent tail of this page, measured by a second instrument: the Global Terrorism Database, 1970–2020, frozen when Washington cut the funding. For every U.S. attack the GTD asked whether it crossed a border — foreign perpetrators, foreign logistics, foreign cause. Where the coders answered, America’s terrorism splits almost exactly down the middle: 748 home-grown, 757 internationally connected. But watch the gray line — after 2009 START largely stopped resolving the question, so the recent era reads “unknown,” not “domestic.” One note on instruments: the ADL feed on this page has no domestic-terrorism category, so the GTD rows are not a re-count of anything above them — they are the missing series, and the lead chart on this page sums the two instruments into a single line for that reason. This section keeps the GTD’s own domestic/international coding as supporting detail. The cross-border half of the story lives under Global Terrorism.
Count incidents and American terrorism is anti-abortion extremists, left-wing militants, white supremacists and Puerto Rican separatists — mostly bombs, mostly at night, mostly against property. Count bodies and the list inverts into a single name. Both lists are true; pick your metric and you pick your narrative, which is exactly why we show both.
The bars above file 1,252 GTD attacks under one “Left Wing” label — the biggest bucket in the American ledger, and the most misleading, because it is not one movement. It is at least five, separated by decades and by body count. Split the monolith and the eras fall out clean: the Marxist bombing wave of the 1970s that mostly killed nobody, the Puerto Rican independence campaign that ran fifteen years, the eco and animal-rights arson of the 2000s with one death in fifty years, the Black nationalist violence of the early 1970s that carries most of the bucket’s bodies, and a modern residue — anarchist, anti-fascist and single-issue — too small to see at the top level. Methodology, stamped: this panel is GTD only, 1970–2020 — the ADL’s own left-wing rows carry no group detail and cannot be split, so they are charted above but not here. Family assignments are hand-made in the open, same as the top-level mapping: Puerto Rican independence counted as its own family, and the GTD’s generic “left-wing” tags filed under Marxist / Revolutionary as the catch-all.
Swastikas on synagogue walls, cemetery desecration, torched pride flags. Property crime, charted separately — because a spray can is not a knife, and a knife is not a flyer.
Slurs, taunts, threatening mail, Zoom-bombings. Most of it is not chargeable, and this site will not inflate it into “crime” — but the tide is real and worth seeing. The 2023 surge follows October 7 and is the largest single-year jump in the dataset.
White supremacist propaganda drops and rallies. This is protected speech and protected assembly — Article VII defends the flyer and Article CXXIII defends the rally, right up until the brick flies. We count it anyway, because legal and harmless are different words. One honesty note: this feed’s propaganda series goes dark after 2023. The ADL’s published reports say 2024 set records, but those rows never landed in the map’s data feed — so the line ends where the data ends, not where the story does.
Everything above this line comes from one instrument — the ADL’s advocacy feed — and this page’s own stamp admits what that means. So here is the second instrument: the FBI’s hate-crime statistics, thirty-five years of agency-reported bias crime, the same federal machinery behind the Crime Watch tab. The two instruments disagree in exactly the ways you’d predict — the FBI counts crimes a prosecutor could charge, the ADL counts incidents a community reports — and both belong on the page, because neither alone deserves your trust. Read the government’s numbers with the government’s caveat attached: reporting is voluntary. Thousands of police agencies file zero hate crimes year after year; entire states have gone functionally dark in past editions; the 2021 switchover to NIBRS scrambled participation mid-series. Every line below is a floor, not a ceiling — and the floor still shows a country whose recorded bias crime has never been higher than this decade.
Raw counts, not per-capita — big states lead partly because they are big. Each bar splits along the same legal line as the rest of the page: the bright segment is crime, the rest is not.
The Global Terrorism Database — 209,706 attacks across fifty years, built by START at the University of Maryland until Washington cut the funding and collection stopped dead. This ledger is frozen at 2020 and will never update; we serve it anyway, because it is still the only place where a pipe bomb in 1975 Manhattan and a truck bomb in 2014 Mosul sit in the same table. The home-grown slice of America’s ledger — the domestic split and the who-and-how — now lives under Domestic Extremism, where it belongs; this tab keeps the world. Two honest holes up front: 1993 is missing entirely — the original collector physically lost that year’s records, first World Trade Center bombing included — and kill counts include the perpetrators.
Two ledgers on very different scales. The world’s line climbs from hundreds of attacks a year to a 2014 peak of nearly 17,000 as ISIS burned across Iraq and Syria. America’s line runs the other way — the busiest terrorism decade on U.S. soil was the 1970s, fourteen hundred bombings by the FALN, the Weather Underground and their cousins, most of them killing nobody. The deaths chart is one vertical wall: a single September morning accounts for three of every four Americans ever killed by terrorism in this database.
Where terrorism actually happens: the Middle East and South Asia carry over half the entire database. North America — the region, not just the U.S. — is under two percent of incidents and under one percent of the dead. The deadliest organizations on earth are not the ones in American nightmares; they are the ones grinding through Baghdad, Kabul and Maiduguri year after year.
The GTD froze at 2020; the counting didn’t stop. GTTAC’s GRID picked up the State Department’s statistical annex where START lost it — same lineage, worldwide coverage, 2018 to now, bulk data free for the taking. But it is a different instrument: in the three years both databases cover, they disagree on the totals. So the lines are charted side by side in the overlap and never summed — a handoff you can inspect, not a splice you have to trust.
The raw influence wire — every Lobbying Disclosure Act filing as it posts to the Senate database. Not annual summaries, not aggregates: this firm registered to lobby for this client on this issue, posted today. Amounts are what the lobbyists themselves reported being paid. Refreshes hourly.
What the Department of Defense has actually spent this fiscal year, title by title, against the same stretch of last year — the real outlays, not the press-release topline. Straight from the Treasury’s Monthly Treasury Statement. Trend line is national defense as a share of the whole economy, back to the Reagan buildup.
Where American troops are permanently stationed, country by country — the Pentagon’s own count, not a think-tank estimate. These are forces with orders to a base abroad; rotational deployments and contingency forces aren’t in here, which is why places that host thousands of rotating troops can still show a small permanent number.
How big the military actually is, by component — active-duty end strength against the ceiling Congress authorized, plus the Guard and Reserve. Every active branch came in at or near its authorized number; a few ran over.
The size of the Navy against the one it’s built to deter. The battle-force count versus the 355-ship floor Congress wrote into law — and versus China’s fleet, now the largest in the world by hull count. The comparison isn’t apples-to-apples: the PLA Navy number folds in smaller coastal combatants the U.S. doesn’t count as battle force.
What the force costs in paychecks — the Military Personnel appropriation, fiscal year to date — and the across-the-board basic-pay raise Congress writes into law each year. The 2025 raise carried an extra targeted bump for the most junior ranks.
Whether the force can fill its own ranks — active-duty enlisted accessions against the goal each service set, fiscal year by fiscal year. FY2023 was the trough, when the services collectively missed by tens of thousands; FY2025 the Pentagon called the best in fifteen years.
Whether the force keeps the people it already has — active-component enlisted reenlistments against goal, summed across each service’s retention zones. Even through the recruiting trough, every service held its reenlistment line.
What the country spends on those who served — Department of Veterans Affairs outlays this fiscal year against last — and how veterans are faring in the labor market, per the Bureau of Labor Statistics.
Follow the money out the other side of the defense budget — the prime-contract dollars the Pentagon obligated to its biggest suppliers, straight from the federal procurement system of record. One company clears more than the next four combined.
The one panel here that is not hard data. There’s no official approval rating for a Secretary of Defense the way there is for a president — only sporadic polls. They’re fenced off here as exactly that: survey snapshots, every point stamped with the pollster, the date, and whether it asked about job approval or personal favorability. Read the trend, not any single number.
The monthly US balance on goods and services — what the tariffs were supposed to close. Each bar is one month in the red; the dashed line is the 2021–24 average. Census & BEA via FRED, seasonally adjusted. Refreshes every six hours.
What crude, natural gas and pump gasoline actually cost — EIA weekly spot and retail prices, a year of history. Refreshes every three hours.
Daily vessel transits through the six lanes that move the world’s trade — satellite AIS counted by the IMF’s PortWatch. When a strait chokes, it shows here first.
The mineral price of the harvest. Monthly world prices for the three nutrients every crop needs — potash, nitrogen and phosphate — set against how much of each America has to import. When a strait chokes or an exporter slams the door, the cost of next year’s food starts here. World Bank Pink Sheet; import reliance from the USGS.
Changes to the Harmonized Tariff Schedule as the USITC publishes them, and the trade deficit they’re supposed to fix — straight from the federal ledgers.
The Treasury’s most recent sanctions action — every name added, removed or amended on the SDN list, and the programs doing the adding. OFAC delta files, published per action day.
Every week of crude oil and fuel prices the EIA has on record — WTI and Brent crude back to 1986, retail gasoline since 1990, diesel since 1994, Henry Hub since 1997. The $147 peak, the COVID crash, the Ukraine spike, all of it.
How many refineries the country operates, how much crude they can run, and how hard they are running it. EIA Refinery Capacity Report, counted as of January 1 every year since 1982, plus the weekly utilization tape.
Operable crude distillation capacity by refining district and by state — who actually processes the barrels, and how concentrated it is.
Weekly US crude oil production since 1983 and the share coming from tight oil (shale / fracking) since 2000. EIA Weekly Petroleum Status Report and Drilling Productivity Report.
Crude oil and total petroleum products — what crosses the border each week, in thousands of barrels per day. The US became a net petroleum exporter in 2020.
Petroleum products supplied to the US market each week — the closest proxy for actual consumption. Total and broken down by gasoline, distillate (diesel and heating oil), and jet fuel.
Every year since 1880, how far global land-ocean temperature ran from the 20th-century average — NOAA NCEI, one number per year. The zero line is the 1901–2000 mean.
Atmospheric CO₂ measured on the slope of Mauna Loa since 1958 — the Keeling Curve, monthly means from NOAA’s Global Monitoring Laboratory.
Weather and climate disasters costing $1 billion or more, CPI-adjusted and counted by NOAA NCEI since 1980. The count and the bill, per year.
Methane traps ~80x more heat than CO₂ over 20 years and nobody talks about it — globally averaged marine surface CH₂₄, monthly means since 1983, NOAA Global Monitoring Laboratory.
US energy CO₂ emissions peaked in 2007–08 and have fallen ever since — no mandate did that; fracked gas outcompeted coal. Meanwhile the Keeling Curve above never blinked. EIA Total Energy, monthly since 1973.
Arctic sea ice extent by day of year — this year’s line against the 1980s average and the 2012 record low. NSIDC Sea Ice Index, daily.
Four American tide gauges, half a century each — monthly mean sea level averaged to annual, shown in inches above each station’s mid-1970s baseline. NOAA CO-OPS.
Every wildfire incident reported to the interagency system this year — count, acreage, and what’s burning right now. NIFC/WFIGS, refreshed hourly.
Federal disaster declarations by year — every county-level designation under FEMA’s major-disaster, emergency and fire programs. The billion-dollar chart above is the bill; this is the frequency.
How much of the lower 48 is abnormally dry or worse — the U.S. Drought Monitor’s weekly D0–D4 area percentages, two years of history.
The IMF’s global food price index — what a basket of food commodities costs the world, monthly, ten years of history via FRED.
What corn, wheat, soybeans and cattle actually trade for — front-month CME futures, a year of weekly closes.
Where the majors are trading right now — S&P 500, Dow, Nasdaq, VIX, the 10-year, and Bitcoin, with the day’s intraday path in five-minute bars.
The longer arc — daily closes for the last six months.
The high-yield spread — the extra yield junk-rated borrowers pay over Treasuries. Calm markets sit near 3; panics have historically driven it past 10, and 2008 pushed it past 20. FRED now carries roughly three years of the ICE index, so that is the window shown — no pretend depth.
The resident analyst’s current stock, crypto and ETF picks, plus the graded track record — biggest winners and losers included, nothing cherry-picked.
Every edge the model likes today — pick, price, and the edge it thinks it has — plus the full graded record since inception.
The board — scheduled games with consensus moneylines, spreads and totals across the books the model compares.
How it actually went — final scores from yesterday across the majors, and how yesterday’s bets graded out.
Sources: U.S. Treasury “Debt to the Penny”, Federal Reserve Economic Data (FRED),
Real-Time Crime Index, WARN Firehose, European Central Bank reference rates, and exchange quotes.
Denominators are official estimates; the per-taxpayer figure assumes the IRS filing rate.
Waiting-list ledger: CMS Provider Data Catalog — Timely and Effective Care at hospital, state, national and Veterans Health Administration level (measures OP_22 “left before being seen” and OP_18a/b/c/d emergency-department throughput) — plus CDC National Health Interview Survey quarterly early-release and historical cost-barrier tables. The Canadian comparison figure is SecondStreet.org’s FOI compilation, not a government series, and is labelled as such on the page.
Health surveillance: CDC National Wastewater Surveillance System, NNDSS weekly notifiable-disease tables, NCHS provisional death counts, NHSN weekly hospital admissions, BLS Consumer Price Index (medical care & health insurance), and CMS marketplace effectuated-enrollment data.
Global health: WHO Disease Outbreak News, WHO FluNet influenza surveillance, and GDACS disaster alerts.
Disasters: FEMA OpenFEMA declarations, NWS active alerts, USGS earthquake feed. Border: CBP nationwide encounters. Crime: FBI Crime Data Explorer.
Refining: EIA Refinery Capacity Report (annual, as of January 1) and Weekly Petroleum Status Report utilization and crude runs.
Trade & energy: EIA weekly prices, IMF PortWatch chokepoint transits, USITC Harmonized Tariff Schedule, BEA trade accounts via FRED, and OFAC sanctions delta files.
Markets & sports: exchange quotes via public chart APIs, the resident analyst\u2019s cached picks and graded record, consensus sportsbook odds, and ESPN scoreboards.
Climate: NOAA NCEI global temperature and billion-dollar disaster records, NOAA GML Mauna Loa CO₂. Agriculture: U.S. Drought Monitor, IMF food price index via FRED, CME futures quotes.