# Statistics Example
Demonstrates the `StatisticsOperators` trait — streaming numeric aggregations
that chain onto any `Stream<T: ToPrimitive>` and emit `f64`. It covers
exponential smoothing (`ewma`, taking an `EwmaSpan` of `PerTick` or `HalfLife`)
and the aggregates `mean`/`variance`/`std`/`sum`/`min`/`max`/`median`. Each
aggregate takes a `Window` — `Count(n)` (the last `n` samples), `Time(duration)`
(the last `duration` of graph time), or `Unbounded` (cumulative, the whole
stream) — so one method serves both the rolling and cumulative cases. The moment
operators additionally take a `Weighting`: `Count` (every sample equal) or `Time`
(each sample weighted by how long it was in effect), so irregular tick spacing is
handled correctly.
One synthetic price stream feeds a spread of statistics, which are `combine`d
into a row, sampled twice a second, and printed live as a table with `for_each`.
## Run
```sh
cargo run --example statistics
```
## Code
```rust
use std::rc::Rc;
use std::time::Duration;
use wingfoil::adapters::statistics::*;
use wingfoil::*;
fn main() {
let price = ticker(Duration::from_millis(100))
.count()
.map(|n: u64| 100.0 + ((n as f64) * 0.6).sin() * 5.0);
// One statistic per column, all derived from the shared price stream.
let labels = ["price", "ewma", "sma", "std", "min", "max", "twap"];
let columns: Vec<Rc<dyn Stream<f64>>> = vec![
price.clone(),
price.ewma(EwmaSpan::PerTick(0.3)),
price.mean(Window::Count(10), Weighting::Count),
price.std(Window::Count(10), Weighting::Count),
price.min(Window::Count(10)),
price.max(Window::Count(10)),
price.mean(Window::Unbounded, Weighting::Time),
];
// Header first, then stream: `combine` collects the columns (in order) into
// one Stream<Burst<f64>> whose value each tick is a row holding every
// column's latest value. Sample that row twice a second and print it live.
print!("{:>6}", "time");
for label in labels {
print!(" {label:>7}");
}
println!();
let trigger = ticker(Duration::from_millis(500));
let table = combine(columns).sample(trigger).for_each(|row, time| {
print!("{:>5.1}s", f64::from(time) / 1e9);
for value in &row {
print!(" {value:>7.2}");
}
println!();
});
Graph::new(vec![table], RunMode::HistoricalFrom(NanoTime::ZERO), RunFor::Duration(Duration::from_secs(5)))
.run()
.unwrap();
}
```
See [`main.rs`](./main.rs) for the full runnable version.
## Output
```
time price ewma sma std min max twap
0.0s 102.82 102.82 102.82 0.00 102.82 102.82 102.82
0.5s 97.79 101.29 102.37 2.70 97.79 104.87 103.29
1.0s 101.56 99.00 99.84 3.73 95.02 104.87 99.96
1.5s 99.13 101.43 100.14 3.75 95.02 104.99 101.10
2.0s 100.17 98.29 99.89 3.78 95.10 104.99 100.00
2.5s 100.54 101.97 100.08 3.80 95.10 104.83 100.63
3.0s 98.77 97.82 99.95 3.81 95.04 104.83 100.03
3.5s 101.91 102.36 100.01 3.82 95.04 105.00 100.42
4.0s 97.46 97.51 100.02 3.82 95.07 105.00 100.06
4.5s 103.12 102.57 99.94 3.81 95.07 104.78 100.29
5.0s 96.36 97.41 100.09 3.79 95.06 104.78 100.09
```