NormalFit

Cargo featurefeatures = ["math"]

View Markdown Open in Claude Open in ChatGPT

Samples pouring into a histogram, then the normal curve that fits them. samples is a slice, and shown (a tracker) says how many of them are binned so far: bins bars over range as a density, no taller than the top of the y range. The curve has the mean and standard deviation of the samples shown and draws with curve_enter. With stats it writes n = 1500 mean 0.03 sd 1.01 at stats_at (pixels). normal_samples(seed, n, mean, sd) gives the same samples for the same seed (the random bits match on every machine, the logarithm and cosine may differ in the last bit); mean_sd and normal_pdf are the numbers behind the curve. Gallery scenes: NormalFit, HistogramToNormal.

NormalFit demo, first frame — `NormalFit` 4.6 s

HistogramToNormal demo, first frame — `HistogramToNormal` 4.6 s

use fframes::animation::KeyFramesAnimation;
use fframes_kit::prelude::*;
use fframes_kit::kits::math::{mean_sd, normal_samples};

fn figure<'a>(
    ui: &mut Ui<'a, '_, '_>,
    samples: &'a [f64],
    shown: &'a KeyFramesAnimation<f32>,
) -> Svgr<'a> {
    let axes = Axes {
        x_range: Range::new(-3.0, 3.0, 1.0),
        y_range: Range::new(0.0, 0.5, 0.1),
        ..ui.defaults()
    };
    let map = axes.map(ui);
    let fit = ui.show(NormalFit {
        samples,
        shown: shown.into(),
        axes: map,
        range: (-3.0, 3.0),
        bins: 24,
        curve_enter: Some(Create { delay: 3.0, ..Default::default() }.into()),
        ..ui.defaults()
    });
    let axes = ui.show(axes);
    svgr!(<g>{axes}{fit}</g>)
}

let samples = normal_samples(7, 1500, 0.0, 1.0);
let (mean, sd) = mean_sd(&samples);
assert!(mean.abs() < 0.1 && (sd - 1.0).abs() < 0.1);

API

Components are plain structs. Set the fields you need and fill the rest with ..ui.defaults(): the defaults below come from the current theme (theme. values are fields of Theme).

NormalFit

Field	Type	Default	Description
`samples`	`&[f64]`	`&[]`	The sample values to bin; only the first `shown` of them count.
`shown`	`ValueTracker`	`0.0.into()`	How many of the samples are shown so far; a tracker, so they can pour in over time.
`axes`	`AxesMap`	`AxesMap::default()`	The axes (an `AxesMap`) the bars and the curve sit on; the bars are no taller than the top of its y range.
`range`	`(f64, f64)`	`(-3.0, 3.0)`	The values `(min, max)` the bars cover, in the axes’ x units.
`bins`	`usize`	`24`	Number of bars the range is cut into; the bar heights are a density.
`fill`	`Color`	`theme.colors.accent`	Fill of the bars.
`curve_color`	`Color`	`theme.colors.highlight`	Colour of the fitted normal curve.
`curve_width`	`f32`	`theme.stroke.thick`	Width of the curve in pixels.
`curve_enter`	`Option<EnterAnim>`	`None`	How the curve appears; `None` shows it from the start.
`stats`	`bool`	`true`	Writes the mean and standard deviation of the samples shown at `stats_at` when `true`.
`stats_at`	`(f32, f32)`	`(1500.0, 300.0)`	Where the statistics are written, in pixels of the canvas.
`stats_enter`	`Option<FadeIn>`	`None`	Fade in of the statistics; `None` shows them from the start.

Feature math · Source