Macro Views
EXPERIMENT Live · rendered from the DB each load · drag to orbit, wheel to zoomThe macro data we actually hold, made tangible. Three experiments, all hand-built three.js over real rows — nothing here is a stock chart library.
Data specimens — one real object from each dataset, as the scenes receive it
stocks_market_regime (665 total, feeds Terrain + Tunnel){
"d": "2026-08-21",
"s": 32,
"l": "risk_on",
"v": 15.1300000000000007815970093361102044582366943359375,
"vp": 9,
"dd": -1.600000000000000088817841970012523233890533447265625
}
{
"tk": "ABEV",
"q": 65,
"v": 43,
"s": 0,
"g": 53,
"t": "gem",
"sec": "Consumer Defensive"
}
How to read this — 60 seconds, no jargon
Ignore which way it's currently spun — the landscape is labeled. Floating signs at the ends of the ridges name them (MOOD, FEAR, PAIN), and the timeline runs between the sign that says ← 2024 and the one that says TODAY →. Every trading day is one thin slice of the terrain — about 650 slices — and the three ridges describe the same days, so their features line up. If you get lost, hit ⟲ reset view (top right of the scene).
- MOOD — the market's daily mood, from our regime engine. Tall + green = confident, risk-taking market. Gray = neutral. Amber = nervous. Sunken + red = stress, the market in fear mode.
- FEAR — the VIX ("Wall Street's fear index") compared to its own history. Taller = more fear than usual. It spikes at exactly the moments MOOD turns red.
- PAIN — how far the S&P 500 sat below its recent high. Level ground = at the highs; a canyon = a correction in progress. The deepest canyon on the map is the April 2025 crash.
The one-sentence version: find a spot where MOOD goes red, FEAR spikes, and PAIN digs a canyon all at once — that's a market storm. The long green plateaus are the calm bull stretches between them. This is the "tide" our stock-picking lenses can't see on their own, drawn as geography.
Yes, it really goes back to January 2024: the regime score is pure math over daily S&P / VIX / Russell closes, and that price history was backfilled — so the terrain is exact reconstruction, not estimation. It grows one slice per trading day from here on.
How to read this — 60 seconds, no jargon
Each glowing hub is a sector (the same families as the GICS categories tree), sized by the total market value of the names we've analyzed in it, and colored by where its money is going — green hub = the sector's giants are trading above their own recent trend, red = below.
- Every star orbiting a hub is one company. Its altitude is the whole story: floating high above the sector's disc = the stock is trading well above its own 50-day average price — buyers have been pushing it up for weeks, money flowing in. Sunk far below the disc = money flowing out.
- Size = company size (a trillion-dollar name is a big planet, a small-cap is a speck). Color = our verdict on it — gold gem, cyan diamond, red avoid, slate watch.
- Hover anything — stars show the exact numbers (vs 50-day, vs 200-day, distance off its high); hubs show the sector totals.
The one-sentence version: sectors whose swarm floats high are where money has been flowing for the past couple of months — and a red-colored star floating high is a stock the market loves but we don't (the PLTR shape), while a gold star sunk low is one we like that the market is still ignoring.
"Trend" here = price vs its own 50-day moving average (a ~2–3 month flow proxy), refreshed by the screener's quote fills — honest, but not literal fund-flow data.