There are 2D, 3D, and nD versions of the ScagnosticsJS library correspondingly.

The paper: ScagnosticsJS: Visual Features for the Web
References:
@inproceedings {s.20201022, booktitle = {Eurographics 2020 - Short Papers}, editor = {Wilkie, Alexander and Banterle, Francesco}, title = {ScagnosticsJS: Extended Scatterplot Visual Features for the Web}, author = {Pham, Vung and Dang, Tommy}, year = {2020}, publisher = {The Eurographics Association}, ISSN = {1017-4656}, ISBN = {978-3-03868-101-4}, DOI = {10.2312/egs.20201022} }
The presentation at The Eurographics 2020 conference:

You can use these playground pages to explore the underlying scagnostics calculation processes and the visualizations of their intermediate results. They contain exemplar scatterplots for each of the 9 scagnostics (outlying, skewed, clumpy, sparse, striated, convex, skinny, stringy, and monotonic) features for you to explore:

2D: https://idatavisualizationlab.github.io/ScagnosticsJS/scagnostics

3D: https://idatavisualizationlab.github.io/ScagnosticsJS/scagnostics3d

nD: https://idatavisualizationlab.github.io/ScagnosticsJS/scagnosticsnd

Figure 1: Scagnostics measures and exemplar plots that they target for 2D (a), 3D (b), and nD (c).

You can download scagnostics.js from here

You can add scagnostics.js to your web page using this code snippet:

<script type="text/javascript" src="https://idatavisualizationlab.github.io/ScagnosticsJS/scagnostics/build/js/scagnostics.min.js"></script>

You will need to pass an array of 2D scatter plot points as [[x1, y1], [x2, y2], [x3, y3], [x4, y4],...[xn, yn]]. If you would like to pass some data (say ids of the points) then you may pass them as data attribute of the point. This data field will then be passed into the result after normalization or binning, so you could use this to get back the original point.

You may use default scagnostics options or you can optionally pass an options object with one or combination of the following fields

• binType: as 'leader' (default) or 'hexagon' to use leader binning or hexagon binning correspondingly.
• startBinGridSize: default is 40, an integer to start the binning process with (it starts with 40x40 bins in this case).
• isNormalized: default is false, if it is true (in case the data is already normalized) then ScagnosticsJS will skip the normalizastion step to speed up the calculation time.
• isBinned: default is false, if it is true (in case the data is already binned) then ScagnosticsJS will skip the binning step to speed up the calculation time.
• outlyingUpperBound: default is undefined, you may specify this outlying upper bound value to decide if a length of the Minimum Spanning Tree, built in the process of scagnostics calculation, is outlying or not.
• minBins: default is 50, this is the minimum number of bins that you expect to have after binning process.
• maxBins: default is 250, this is the maximum number of bins that you expect to have after binning process.

    //Create an array of random 2D points
    let points = [];
    for (let i = 0; i < 100; i++) {
        let x = i* (3*Math.PI/100);
        let y = Math.sin(x);
        points.push([x+Math.random()/100, y+Math.random()/100]);
    }
    //Create scagnostics object
    let scag = new scagnostics(points); 

    //Create an array of 2D random points
    let points = [];
    for (let i = 0; i < 100; i++) {
        let x = i* (3*Math.PI/100);
        let y = Math.sin(x);
        points.push([x+Math.random()/100, y+Math.random()/100]);
    }
    //Set the scagnostics options
    const options = {
        binType: 'leader',
        startBinGridSize: 20,
        isNormalized: false,
        isBinned: false,
        outlyingUpperBound: undefined,
        minBins: 50,
        maxBins: 250
    };
    let scag = new scagnostics(points, options); 

• scag.normalizedPoints returns an array of 2D points after normalized to the range [0, 1]. Each point will contain a data field that passed from the original point (if you assigned one), so you could use to get back the information from the original point.

• scag.bins will return all the bins. A bin contains an array of points belonging to that bin and x, y attributes for the location (center) of the bin.

• scag.outlyingScore returns the scagnostics outlying score.

• scag.skewedScore returns the scagnostics skewed score.

• scag.clumpyScore returns the scagnostics clumpy score.

• scag.sparseScore returns the scagnostics sparse score.

• scag.striatedScore returns the scagnostics striated score.

• scag.convexScore returns the scagnostics convex score.

• scag.skinnyScore returns the scagnostics stringy score.

• scag.stringyScore returns the scagnostics stringy score.

• scag.monotonicScore returns the scagnostics monotonic score.

You can download scagnostics3d.js from here

You can add scagnostics3d.js to your web page using this code snippet:

<script type="text/javascript" src="https://idatavisualizationlab.github.io/ScagnosticsJS/scagnostics3d/build/js/scagnostics3d.min.js"></script>

You will need to pass an array of 3D scatter plot points as [[x1, y1, z1], [x2, y2, z2], [x3, y3, z3], [x4, y4, z4],...[xn, yn, zn]]. If you would like to pass some data (say ids of the points) then you may pass them as data attribute of the point. This data field will then be passed into the result after normalization or binning, so you could use this to get back the original point.

You may use default scagnostics options or you can optionally pass an options object with one or combination of the following fields

• binType: as 'leader' (default), currently only 'leader' binning is supported.
• startBinGridSize: default is 40, an integer to start the binning process with (it starts with 40x40x40 bins in this case).
• isNormalized: default is false, if it is true (in case the data is already normalized) then ScagnosticsJS will skip the normalizastion step to speed up the calculation time.
• isBinned: default is false, if it is true (in case the data is already binned) then ScagnosticsJS will skip the binning step to speed up the calculation time.
• outlyingUpperBound: default is undefined, you may specify this outlying upper bound value to decide if a length of the Minimum Spanning Tree, built in the process of scagnostics calculation, is outlying or not.
• minBins: default is 30, this is the minimum number of bins that you expect to have after binning process.
• maxBins: default is 200, this is the maximum number of bins that you expect to have after binning process.

    //Create an array of random 3D points
    let noise = ()=> Math.random()/10;
        let points = [];
        for (let x = 0; x < 150; x++) {
            points.push([x + noise(), Math.sin(x * Math.PI / 20) + noise(), Math.cos(x * Math.PI / 20) + noise()])
        };
    //Create scagnostics object
    let scag = new scagnostics3d(points); 

    //Create an array of random 3D points
    let noise = ()=> Math.random()/10;
        let points = [];
        for (let x = 0; x < 150; x++) {
            points.push([x + noise(), Math.sin(x * Math.PI / 20) + noise(), Math.cos(x * Math.PI / 20) + noise()])
        };
    //Set the scagnostics options
    const options = {
        binType: 'leader',
        startBinGridSize: 20,
        isNormalized: false,
        isBinned: false,
        outlyingUpperBound: undefined,
        minBins: 30,
        maxBins: 200
    };
    let scag = new scagnostics3d(points, options); 

• scag.normalizedPoints returns an array of 3d points after normalized to the range [0, 1]. Each point will contain a data field that passed from the original point (if you assigned one), so you could use to get back the information from the original point.

• scag.bins will return all the bins. A bin contains an array of points belonging to that bin and x, y, z attributes for the location (center) of the bin.

• scag.outlyingScore returns the scagnostics outlying score.

• scag.skewedScore returns the scagnostics skewed score.

• scag.clumpyScore returns the scagnostics clumpy score.

• scag.sparseScore returns the scagnostics sparse score.

• scag.striatedScore returns the scagnostics striated score.

• scag.convexScore returns the scagnostics convex score.

• scag.skinnyScore returns the scagnostics stringy score.

• scag.stringyScore returns the scagnostics stringy score.

• scag.monotonicScore returns the scagnostics monotonic score.

You can download scagnosticsnd.js from here

You can add scagnosticsnd.js to your web page using this code snippet:

<script type="text/javascript" src="https://idatavisualizationlab.github.io/ScagnosticsJS/scagnosticsnd/build/js/scagnosticsnd.min.js"></script>

You will need to pass an array of nD scatter plot points as [[x1, y1, z1, ...], [x2, y2, z2, ...], [x3, y3, z3, ...], [x4, y4, z4, ...], ... [xn, yn, zn, ...]]. If you would like to pass some data (say ids of the points) then you may pass them as data attribute of the point. This data field will then be passed into the result after normalization or binning, so you could use this to get back the original point.

You may use default scagnostics options or you can optionally pass an options object with one or combination of the following fields

• binType: as 'leader' (default), currently only leader binning is supported.
• startBinGridSize: default is 20, an integer to start the binning process with (it starts with 20x20x20... bins in this case). For high number of dimensions, the number of bins to start with increases exponentially, so consider reducing this number for better performance.
• isNormalized: default is false, if it is true (in case the data is already normalized) then ScagnosticsJS will skip the normalizastion step to speed up the calculation time.
• isBinned: default is false, if it is true (in case the data is already binned) then ScagnosticsJS will skip the binning step to speed up the calculation time.
• outlyingUpperBound: default is undefined, you may specify this outlying upper bound value to decide if a length of the Minimum Spanning Tree, built in the process of scagnostics calculation, is outlying or not.
• minBins: default is 30, this is the minimum number of bins that you expect to have after binning process.
• maxBins: default is 200, this is the maximum number of bins that you expect to have after binning process.
• outlyingCoefficient: default is 3, this is used to calculate the outlying threshold on the MST lengths using this formula: upperBound = 75th quantile + outlyingCoefficient*IQR. We provide this option since in high dimensional data the default formula upperBound = 75th quantile + 1.5*IQR may not work in detecting outliers.
• incrementA: default is 2, in high dimension data, number of bins will change dramatically if we change the bin grid size, therefore, we can use the formula binSize = incrementA*binSize + incrementB to control the increment of the binSize.
• incrementB: default is 0, explained in incrementA section.
• decrementA: default is 1/2, in high dimension data, number of bins will change dramatically if we change the bin grid size, therefore, we can use the formula binSize = decrementA*binSize + decrementB to control the decrement of the binSize.
• decrementB: default is 0, explained in incrementA section.
• distanceWeights: default is 1 for every dimension. In high dimensional data, we may need the option to control the weights of each dimension in the contribution to the distance calculation among points. In this case we can specify this as an array, a value in range [0, 1] represents the weight of the corresponding dimension in distance calculation. E.g., [0.5, 1, 1, 1, 1, 1, 0.5] means that there are 7 dimensions and the first and last have the weights of 0.5 in calculating distances.

    /***********RANDOM DATA*******************/
    let random = Math.random,
        points = [];
    //100 data points
    for (let i = 0; i < 100; i++) {
        //each point of 10 dimensions
        let point = [];
        for (let j = 0; j < 10; j++) {
            point.push(random());
        }
        points.push(point);
    }
    //Create scagnostics object
    let scag = new scagnosticsnd(points); 

    /***********RANDOM DATA*******************/
    let random = Math.random,
        points = [];
    //100 data points
    for (let i = 0; i < 100; i++) {
        //each point of 10 dimensions
        let point = [];
        for (let j = 0; j < 10; j++) {
            point.push(random());
        }
        points.push(point);
    }
    //Create options
    let options = {
        startBinGridSize: 30,
        minBins: 30,
        maxBins: 100,
        outlyingCoefficient: 0.5,
        incrementA: 1,
        incrementB: 5,
        decrementA: 0.9,
        decrementB: 0,
        distanceWeights: [0.5, 0.5, 0.5, 1, 1, 1, 1, 1, 1, 0.5]
    };
    let scag = new scagnosticsnd(points, options); 

• scag.normalizedPoints returns an array of nD points after normalized to the range [0, 1]. Each point will contain a data field that passed from the original point (if you assigned one), so you could use to get back the information from the original point.

• scag.bins will return all the bins. A bin contains an array of points belonging to that bin and x, y, z, ... attributes for the location (center) of the bin.

• scag.outlyingScore returns the scagnostics outlying score.

• scag.skewedScore returns the scagnostics skewed score.

• scag.clumpyScore returns the scagnostics clumpy score.

• scag.sparseScore returns the scagnostics sparse score.

• scag.striatedScore returns undefined in this current version.

• scag.convexScore returns undefined in this current version.

• scag.skinnyScore returns undefined in this current version.

• scag.stringyScore returns the scagnostics stringy score.

• scag.monotonicScore returns the scagnostics monotonic score.