The S&P 500 index
The whole index, read the way I actually read it: built from source data, refreshed on demand.
Every rectangle is one company, sized by its share of the index and coloured by its last month.
Each rectangle is one company, and its area is how much of the index that company actually is. The handful of giants across the top are the ones that decide where the index goes. The hundreds of slivers at the edges barely register, no matter what they do. Use the dropdown to isolate a single sector, or switch the colour to see sectors instead of moves.
Every cell is one month since 1928: red lost, green gained. The average month is quietly, boringly positive.
Read left to right for one year, top to bottom for a century. Each square is a single month, and the stronger the colour the bigger the move. The last column is the year as a whole. Scan for the dark red bands and you have found 1929, 1987, 2008 and 2020 without being told where to look. Then notice how much of the wall is faint green: most months are small gains, and that is the whole engine.
A rising level can hide a narrowing market. These four are the health checks.
Both lines share the same clock: when the green share falls while the index climbs, fewer companies are carrying the rally.
The top line is the index itself. The green area below counts how many of the 500 are still above their own average price of the last 200 trading days, a rough test of "still trending up". When both move together the rise is broad and healthy. When the index keeps climbing while the green share falls away, a shrinking handful of giants is holding the whole thing up, and that narrowing has come before most of the ugly turns. Around 70% is broad; below 40% while the index is near a high is the warning shape.
Every dip is how far the index sat below its last record. Drag to zoom, double-click to reset.
The line touches the top only when the market sets a fresh record. The deep valleys are the crashes, and their width is the part people forget: getting back to even took years, not weeks.
Every day since 2006, bucketed by size of move; the curve is what a textbook predicts.
The bars are what actually happened, the curve is the theory. Look at the far left and right: real life pokes out well beyond the curve. Those are the crashes and melt-ups, and they turn up far more often than the textbook allows, which is why "once in a century" seems to happen every decade.
Up is what it paid, right is how hard it shook; bubble size is index share, colour is sector. Pick a sector anywhere and the rest step back.
Top left is the prize: a good return without the white knuckles. Bottom right is the trap: all of the stomach churn and none of the payoff. Most things sit on the diagonal in between, because you are usually paid for taking risk, not rewarded for avoiding it.
Eleven sectors, eleven very different personalities. Pick one anywhere, from the dropdown above or any legend, and the whole page follows it.
Rank 1 is the year's best sector since 1999; hover a line to follow it.
Each line is one sector, and its height is where that sector finished the year against the other ten. Lines crossing means leadership changed hands. A line that hugs the top for years is a sector that led a whole era, and a plunge from top to bottom is a boom that ended.
Each band is a sector's slice of the index's total value.
The thickness of a band is how much of the whole index that sector is worth. Bands that fatten over time are sectors quietly taking over the market, and bands that thin are losing their grip. Technology swelling at the expense of everything else is the story of the last decade.
Ten years of monthly moves, with like-behaving sectors grouped side by side.
Find a row, run across to a column, and the square where they meet says how closely those two sectors rise and fall together. Green squares are pairs that move as one, which means owning both protects you less than it looks. Pale red squares are pairs that go their own way, which is where real diversification lives.
Calmest sector first; the grey core repeats it in every panel, so the colour is that sector's extra.
Each hill is the shape of a typical day for that sector. A tall narrow hill means almost every day is a small move. A low, wide hill means that sector regularly lurches in both directions. Every panel is drawn on exactly the same scale, so wider really does mean wilder.
Five traits scored against the other ten sectors; further out means more.
The shape is the sector's fingerprint. Value means it looks cheap against its earnings. Momentum means it has been running hot lately. Low volatility means it moves gently. Yield means it pays big dividends. Quality means steady, profitable businesses. A spike toward one corner is what that sector is known for.
Each sector's push on the index last month; the gold bar is where they all netted out.
The bars are stacked end to end, so each one starts where the last finished and you can watch the index climb or slide as you read down. A big sector that moved a little can outweigh a small sector that moved a lot, which is why the order here is rarely the order of best performers.
The shape was above; here is the substance. Search all 500 names, sort them, star a shortlist.
| Company | Sector | Industry | Weight ▼ | 1-month | Contribution |
|---|
Weight is the company's share of the total; contribution is weight times its month. Stars stay in this browser.
One bubble per industry: swing across, return up, size is index share, colour is sector.
Same idea as the risk and reward chart earlier, but grouped into the trades companies are actually in rather than the company names. Top left is a lot of return for little drama. The big bubbles are the industries with enough weight to drag the whole index with them.
Sectors, then industries, then companies; slice width is share of the index. Click a sector to zoom, the centre to reset.
Rings work outward: the eleven sectors innermost, then the industries inside each, then the companies themselves. How wide a slice is around its ring is how much of the index it accounts for. It is the same information as the map at the top of the page, wrapped into a circle so you can see the layers nest.
No invented numbers: every figure traces to public source data through a pipeline I wrote.
A personal tool I built to understand the S&P 500 from the ground up, then rebuilt for the web. It is a reading of the index, not a trading product and not investment advice. I made it because I wanted to see the market's shape myself rather than take someone else's chart for it.
The membership list comes from the public S&P 500 constituents record on Wikipedia. Prices, market caps and the forward valuation figures come from Yahoo Finance. Both are public sources. The century heatmap uses the index's own monthly closes running back to 1928. The live number at the very top is a convenience quote pulled from Yahoo when the page loads.
Weight is each member's share of the summed market capitalisation of all members, a float-unadjusted proxy for the official index weights. One-month return is the trailing one-month price change. Contribution is start-of-period weight times return, so the eleven sector contributions add up to the index's move. Sectors and industries follow the standard GICS-style groupings carried in the source data.
500 companies across 11 sectors and 112 industries, with data as of the build date. The constituent detail uses current members only, so the sector and industry history does not correct for names that joined or left the index over time.
Market-cap-proxy weights differ slightly from the official float-adjusted index, so totals will not tie out to the last decimal against S&P's published figures. Prices are close-only, not intraday. The live top-line quote is read on load and is not stored anywhere. And to say it plainly: none of this is investment advice.
The page is regenerated from a fresh data pull on demand. The "as of" date is when this particular build was cut. A short fingerprint of the underlying figures is embedded in the page's metadata for provenance, so a given build can be tied back to the exact numbers it was made from.
Built by Dhruv Mandavkar, data as of the build date. An independent personal reading, not affiliated with or endorsed by S&P Dow Jones Indices or MSCI. Companion piece: the live macro signals dashboard.