The quant toolkit for streaming data. How momentum ratios, streak counts, and RSI turn raw playcounts into signals.
Analyst tools, July 29, 2026, 5 min read
A growth percent tells you one thing: is a number bigger than it was before, and by how much. It says nothing about whether that change is speeding up or slowing down.
Two tracks can each post the same 15% jump in monthly streams and be headed in completely opposite directions. One is accelerating into a real moment. The other is coasting on the tail end of a spike that already peaked.
Telling those two apart takes instruments that measure the change in the rate of change, not just the rate itself. That's different math than a simple percent change. It's math borrowed from a different field entirely.
StreamClout borrows three of those instruments from finance: momentum, streaks, and RSI. Streams behave a lot like trading volume, spiking and fading and trending the same way volume does. The tools built to read volume data turn out to work here too.
None of the three are exotic. Each one is a small, well-defined calculation run on daily numbers you already see on every album page, and you can check StreamClout's math against your own by hand if you want to.
The momentum ratio answers one narrow question: is the last week moving faster or slower than the month around it. StreamClout computes it as the average streams per day over the trailing 7 days divided by the average streams per day over the trailing 30 days.
A ratio of exactly 1.0x means the last week looked just like an average week from the last month, cruising at the same pace. Anything meaningfully above or below that says the pace has actually shifted.
There's a subtlety worth knowing before you lean on the number. The 7 day window sits entirely inside the 30 day window, which means those same 7 days are baked into the 30 day average too.
That overlap works against sharp moves. A real spike in the last week drags its own denominator up along with it, which pulls the ratio back toward 1.0x and understates how sharp the move actually was.
Picture a track running 100k streams a day for most of a month that suddenly jumps to 300k a day for the final week. The 30 day average climbs to roughly 146k a day once that hot week is folded in, so the momentum ratio lands near 2.05x rather than the 3.0x you'd get from comparing the hot week to the cold weeks alone.
That's not a bug in the math. It's just a reminder that momentum is a conservative read on acceleration, not an exact multiplier on how much faster things got.
A streak counts consecutive days where the daily stream number sits on one side of the period's own average, above or below. StreamClout walks backward from the most recent day and keeps counting until a day breaks the pattern.
A single spike day produces a 1 day streak. The day right before it and the day right after it usually sit back under the average, so the streak resets almost as fast as it started.
A 5 day streak above average is a different animal. That's five straight days of sustained pressure, and a pattern that consistent is far harder to fake than one loud day.
The strongest simple pattern combines the two signals. A long above-average streak paired with a rising average isn't one lucky break. It's a track building real, compounding momentum day after day.
Watch what happens to that pattern over time and it explains itself. Each above-average day nudges the period average up a little, which raises the bar the next day has to clear to keep the streak alive.
A track that keeps clearing a rising bar for five or six days straight is doing something a single viral clip can't fake. That's demand showing up again and again, not once.
RSI comes straight from technical trading, where it measures whether recent price moves have leaned mostly toward gains or mostly toward losses. StreamClout runs the same math on daily stream deltas and calls it the Heat Index, a score from 0 to 100.
It looks at the day to day changes in the daily stream count, splits the recent run into gains and losses, and expresses the balance as a single number. Above 70 counts as overheated. Below 30 counts as washed out.
Here's the part people get wrong: RSI does not measure size. A track adding a modest number of streams a day can still read overheated, as long as nearly every recent day beat the day before it.
What RSI actually measures is one-sidedness. A steady climb with the occasional small pullback sits comfortably in the middle of the range, while a run of days that are almost all up, or almost all down, pushes the number toward an extreme.
An overheated reading above 70 is worth a second look, not an automatic buy signal. It's telling you the recent run has been unusually one-directional, and one-directional runs are exactly the kind of thing that tends to cool off.
A washed out reading below 30 works the same way in reverse. It says the recent days have leaned hard toward declines, which is often the setup right before a track stabilizes rather than a sign it's dying for good.
None of these three signals predicts anything. They describe the present, not the future.
Extremes fade. They don't guarantee a reversal
Momentum, streaks, and the Heat Index all tend to pull back toward normal on average, but that's a tendency, not a rule. A momentum reading of 2.0x or a Heat Index of 90 doesn't promise a pullback next week. It just says the odds lean that way.
The reason extremes fade is usually mechanical, not mysterious. Playlist placements end, viral clips stop circulating, and press cycles run their course.
Once the thing that caused the spike goes away, the numbers drift back toward whatever baseline the artist was actually running at before it. That's mean reversion in practice, and it's exactly why these three tools describe rather than predict.
That's worth remembering the next time an overheated reading looks like a reason to panic, or a cooling momentum ratio looks like a reason to write an artist off. Neither one is a verdict.
These three signals sit together on every album page, updated from the same daily data. Go see all three computed live and watch how they move against each other on a real catalog.