

Network Chart

A network graph is designed to allow the visualization of relational information. ChartFactor has two types of network charts: one that comes in the standard library called Graph and one more flexible and sofisticated called Network that comes in its own package.

Comparison between the Graph Chart and the Network Chart

Feature	Graph Chart	Network Chart
Simple Graph Visualization	Yes	Yes
Directed Graph	No	Yes
Nodes Custom Styles	No	Yes
Edges Custom Styles	No	Yes
Static layout (specific x and y)	Yes	Yes
Force Layout	Yes	Yes
Circle Layout	No	Yes
Grid Layout	No	Yes
CoSE Layout	No	Yes
Dagre Layout	No	Yes
Spread Layout	No	Yes
Concentric Layout	No	Yes
Custom event listeners	No	Yes

Creating a Basic Graph Chart with local data

In this example we will create a Graph chart and feed it with static data. Note that only edge information is required to render the graph.

var data = {
    'nodes': [
        {
            'color': '#4f19c7',
            'label': 'jquery',
            'y': -404.26147,
            'x': -739.36383,
            'id': 'jquery',
            'value': 4.7252817
        },
        ...
        {
            'color': '#c71969',
            'label': 'backbone',
            'y': -862.7517,
            'x': -134.2215,
            'id': 'backbone',
            'value': 6.1554675
    }],
    'edges': [
        {
            'source': 'jquery',
            'target': 'jsdom',
            'value': 1
        },
        ...
        {
            'source': 'jquery',
            'target': 'xmlhttprequest',
            'value': 1
    }]
}

var edgesInfo = {
    'width': 0.6,
    'curveness': 0.2,
    'opacity': 0.7
};
var nodesInfo = {
    'min': 0.1,
    'max': 60
};

var chart = cf
    .create()
    .graph('Graph')
    .element('chart')
    .set('layout', 'force')
    .set('loadIndicator', 'graph-loading')
    .set('layout', 'force')
    .set('edgesInfo', edgesInfo)
    .set('nodesInfo', nodesInfo)
    .data(data);

chart.execute();

The previous code will render the Graph Chart below:

Creating a Graph chart with provider data

It is recommended before reading this section, to read the documentation on Multiple Queries (click here). To support graph charts, we make use of Chartfactor's multiple query capabilities as we show below:

var edgesInfo = {
    width: 0.6,
    curveness: 0.2,
    opacity: 0.7
};
var nodesInfo = {
    min: 0.1,
    max: 60
};

let metric = cf.Metric('size', 'sum');
let metrics = [metric, cf.Metric('count')];
let grp = cf.Attribute('id').limit(1200).sort('asc', metrics[0]);

let nodesHandler = function (result) {
    result.data.nodes.forEach(n => {
        n.id = n.group[0];
        n.value = n.current.metrics[metric._name][metric._func];
    });
    result.keep = true;
    return result;
};

let edgesHandler = function (result) {
    result.data.edges.forEach(e => {
        e.source = e.group[0];
        e.target = e.group[1];
        e.value = e.current.metrics[metric._name][metric._func];
    });
    result.keep = true;
    return result;
};

let myChart = cf
    .create('nodes')
    .provider('Elasticsearch')
    .source('npmdepgraph-nodes')
    .processWith(nodesHandler.bind(this))
    .groupby(grp)
    .metrics(...metrics)
    .create('edges')
    .provider('Elasticsearch')
    .source('npmdepgraph-links')
    .processWith(edgesHandler.bind(this))
    .rows('source', 'target')
    .limit(500)                // 1000 by default
    .metrics(...metrics)
    .graph('Graph')
    .set('layout', 'force')
    .set('edgesInfo', edgesInfo)
    .set('nodesInfo', nodesInfo)
    .element('chart')
    .execute();

The nodes query in the example above is obtaining the botom 1200 ids by size. Here is the code used when declaring the attribute to group by: let grp = cf.Attribute('id').limit(1200).sort('asc', metrics[0]);.

The edges query obtains the relation betwen nodes. It is using a pivot-style query grouping by source and target, as specified in the rows() function above. Here is the code used: .rows('source', 'target'). The default limit for a rows() query is 1000. You can override this limit by using the limit() function. In the example above, the limit is set to 500.

Note

It is mandatory that the query to retrieve the node is called "nodes" and the query to retrieve edges is called "edges". If you want to see an example of this code in action click here.

Creating a Network Chart with static data

In this example we will create a Network chart and feed it with static data. Note that only edge information is required to render the network. Node information is optional.

    var data = {
        Edges: [ /* is mandatory call it "Edges" and each object have to have a source and a target. */
            {source: 'a', target: 'b'},
            {source: 'a', target: 'c'},
            {source: 'a', target: 'd'},
            {source: 'c', target: 'e'},
            {source: 'c', target: 'f'}
        ],
        Nodes: [] /* is mandatory call it "Nodes" and if no extra information is needed for the node (size, color, image background, etc) the array could by empty */
    }

    cf.create()
    .graph('Network')  // Setting the visual type
    .element('chart')  // Setting the id of the DOM element in where should be render the Network
    .set('layout', {name: 'dagre'}) // Setting the Network layout (see layout section)
    .data(data)        // Injecting static data
    .execute();        // Executing the visual

The previous code will render the Network Chart below:

Adding custom styles

The next example uses the same data as the previous example. In addition, it defines a style array to be assigned to the network chart. Each object in the style array must have a selector attribute (node, edge, or id) and a style attribute containing css-like styles to be applied to network chart.

    var style = [
        {selector: 'node',
            style: {
                'label': 'data(id)',
                'background-color': '#4292B3'
            }
        }, {selector: 'edge',
            style: {
                'line-color': '#4292B3'
            }
        }
    ];
    cf.create()
    .graph('Network')
    .element('chart')
    .set('layout', {name: 'dagre'})
    .set('style', style)  // Setting the style to the Network chart.
    .data(data)
    .execute();

The previous code will render the Network Chart below:

Using the data to style nodes and edges

We can use properties of the data to specify styles of nodes and edges. Use the syntax 'data(propertyName)' to assign the value of that property to a style object. Let's modify our previous example to use this feature:

    var style = [
        {selector: 'node',
            style: {
                'label': 'data(id)',  // Using the id of the node as label
                'background-color': 'data(color)', /* Using the color property of the data object like the background color of the node */
                'height': 'data(size)', // Using size for height
                'width': 'data(size)'   // Using size for width
            } 
        }, {selector: 'edge',
            style: {
                'line-color': 'data(color)',  // Using the property color of the edge data object as line-color
                'curve-style': 'unbundled-bezier'  // This make the line curved
            }
        }
    ];
    var data = {
        /* Now we will modify a little bit the edges data objects to store a color for each edge */
        Edges: [
            {source: 'a', target: 'b', color: 'red'},
            {source: 'a', target: 'c', color: 'red'},
            {source: 'a', target: 'd', color: 'red'},
            {source: 'c', target: 'e', color: 'green'},
            {source: 'c', target: 'f', color: 'green'}
        ],
        /* And modify the nodes data objects to store a color and size for each node */
        Nodes: [
            {id: 'a', size: 10, color: 'blue'},
            {id: 'b', size: 20, color: 'yellow'},
            {id: 'c', size: 15, color: 'black'},
            {id: 'd', size: 30, color: 'orange'},
            {id: 'e', size: 40, color: 'purple'},
            {id: 'f', size: 50, color: 'gray'}
        ]
    };

    cf.create()
    .graph('Network')
    .element('chart')
    .set('layout', {name: 'dagre'})
    .set('style', style)
    .data(data)
    .execute();

The previous code will render the Network Chart below:

Setting arrows to edges

To set arrows on the edges we can modify the properties \<position>-arrow-shape, \<position>-arrow-color, \<position>-arrow-fill and \<position>-arrow-scale.

Valid values for \<position> are:    source : Pointing towards the source node, at the end of the edge.    mid-source : Pointing towards the source node, at the middle of the edge.    target : Pointing towards the target node, at the end of the edge.    mid-target: Pointing towards the target node, at the middle of the edge.

Valid shapes are:    triangle    triangle-tee    triangle-cross    triangle-backcurve    vee    tee    square    circle    diamond    none

    var style = [
        {selector: 'node',
            style: {
                'label': 'data(id)',
                'background-color': 'data(color)',
                'height': 'data(size)',
                'width': 'data(size)'
            }
        }, {selector: 'edge',
            style: {
                'line-color': 'data(color)',
                'curve-style': 'unbundled-bezier',
                'target-arrow-shape': 'triangle',
                'target-arrow-color': 'black'
            }
        }
    ];

Images in Nodes

To put an image in a node you must modify the properties 'background-image' and 'background-fit' in the style of the node as well as specify the path of the image in the data of the node or create a custom selector using the id of the node to specify the path of the background image as shown below:

    var style = [
        {selector: 'node',
            style: {
                'label': 'data(label)',
                'background-color': 'data(color)',
                'height': 'data(size)',
                'width': 'data(size)',
                'background-image': 'data(backgroundImage)',
                'background-fit': 'cover',
                'border-color': '#000',
                'border-width': 1,
                'border-opacity': 0.5
            }
        }, {selector: '#a',
            style: {
                'background-image': 'https://farm8.staticflickr.com/7272/7633179468_3e19e45a0c_b.jpg'
            }
        }, {selector: 'edge',
            style: {
                'line-color': 'black',
                'target-arrow-shape': 'triangle',
                'target-arrow-color': 'black'
            }
        }
    ];
    var data = {
        Edges: [
            {source: 'a', target: 'b', color: 'red'},
            {source: 'a', target: 'c', color: 'red'}
        ],
        Nodes: [{
            id: 'a',
            label: 'bird',
            size: 60,
            color: 'blue'
        }, {
            id: 'b',
            size: 60,
            label: 'cat',
            color: 'yellow',
            backgroundImage: 'https://farm2.staticflickr.com/1261/1413379559_412a540d29_b.jpg'
        }, {
            id: 'c',
            size: 60,
            color: 'black',
            label: 'rose',
            backgroundImage: 'https://farm6.staticflickr.com/5109/5817854163_eaccd688f5_b.jpg'
        }
        ]
    };

Custom Tooltips

Using the same data from the previous example we can develop a small example of how to implement custom tooltips that show the size in addition to the node identifier.

    var tooltipBuilder = function (info) {
        var result = '';
        switch (info.type) {
            case 'node' : {
                result = `<b>Node:</b> ${info.data.id}<br />`;
                result += `<b>Size:</b> ${info.data.size}<br />`;
                break;
            }
            case 'edge' : {
                result = `<b>Edge:</b> ${info.data.source} > ${info.data.target}<br />`;
            }
        }
        return result;
    };
    cf.create()
    .graph('Network')
    .element('chart')
    .set('layout', {name: 'dagre'})
    .set('style', style)
    .set('tooltipBuilder', tooltipBuilder)
    .data(data)
    .execute();

The previous code will render the Tooltips below:

Available layouts

Layout Name	Node Position	Config. Variables	Best Scenarios
preset	preset on data	none	when you have the positions that each node will have
random	random	none	if is a small graph
force	calculated	Edge length, Node spacing	good to detect association and orphan nodes
circle	on the perimeter of a circle	none	if is a small graph
grid	on a grid	spacingFactor	a graph without many connections
cose	calculated	idealEdgeLength, nodeOverlap, nodeRepulsion, edgeElasticity, gravity, etc.	a graph without many connections
dagre	calculated	none	non-cyclic graph
spread	calculated	minDist	a graph without many connections
concentric	calculated	concentric (func), levelWidth (func)	graphs in which few nodes have relationships with many nodes

In this demonstration you can try each of the layouts. (click here)

Registering a click listener

Setting a listener to an event in a graph is done in a similar way as to any other chartfactor visualization. Here is an example:

    var clickListener = function (event) {
        alert(event.data.id);
    };
    var chart = cf.create()
    .graph('Network')
    .element('chart')
    .set('layout', {name: 'dagre'})
    .set('style', style)
    .data(data);

    chart.on('click', clickListener);
    chart.execute();

Creating a Network chart with provider data

It is recommended before reading this section, to read the documentation on Multiple Queries (click here). To support network charts, we make use of Chartfactor's multiple query capabilities as shown below:

    var nodesHandler = function (result) {
        // The callback to be used when executing the query to get the nodes

        // In this example we set custom color for all nodes.
        result.data.Nodes.forEach(x => {
            x.color = '#4292B3';
        });

        // The data should be persistent (important)
        result.keep = true;
        return result;
    };

    var edgesHandler = function (result) {
        // The callback to be used when executing the query to get the edges

        // In this example we set a filter for get only the nodes info that is needed, 
        // and a customs color for the edge line.
        let nodes = [];
        result.data.Edges.forEach(x => {
            if (nodes.indexOf(x.source) === -1) nodes.push(x.source);
            if (nodes.indexOf(x.target) === -1) nodes.push(x.target);
            x.color = '#4292B3';
        });

        result.filters = (nodes && nodes.length > 0) ? [cf.Filter('id').values(nodes)] : [];

        return result;
    };

    cf.create('Edges')        // Query to retrieve the edges info.
    .provider('ProviderName')
    .source('edgesSourceName')
    .fields('source', 'target')
    .limit(1000)              // 100 by default
    .processWith(edgesHandler.bind(this))
    .create('Nodes')          // Query to retrieve the nodes info.
    .provider('ProviderName')   // Could be a different provider.
    .source('nodesSourceName')
    .fields('id', 'value', 'y', 'x', 'label', 'color')    // All this attributes will be added to the node data.
    .limit(1000)              // 100 by default
    .processWith(nodesHandler.bind(this))
    .element('chart')
    .graph('Network')

The edges query obtains the relation betwen nodes. The example above is using a raw data query to retrieve source and target columns, as specified in the fields() function. Here is the code used: .fields('source', 'target'). The default limit for raw data queries is 100. You can override this limit by using the limit() function. In the example above, the limit is set to 1000.

The nodes query in the example above is also using a raw data query to retrieve id, value, y, x, label, and color fields. Same as with the edges query, the default limit for raw data queries is 100. You can override this limit by using the limit() function. In the example above, the limit is set to 1000.

Note

It is mandatory that the query to retrieve the node is called "Nodes" and the query to retrieve edges is called "Edges". If you want to see an example of this code in action click here.

Inject urls of dependency libraries:

In order to not increase the size of the library with the dependencies of an uncommon visualization, by default, ChartFactor does not load the dependencies of the network graph. These are loaded when the graph is instantiated and are left in cache. These are loaded from the official CDN but if you do not have extranet access or for reasons of speed you want to serve them on your own servers, you can inject the addresses of these libraries and they will be loaded in the following way. Example:

    <script>
        window.colaUrl = 'https://chartfactor.com/wp-content/lib/cola.min.js';
    </script>
    <script src="../../lib/CFToolkit.min.js"></script>

Note

It is recommended to define global variables to inject the dependencies before loading the library. The following table shows all the dependencies that can be injected

Vairable	Default Url
colaUrl	http://marvl.infotech.monash.edu/webcola/cola.min.js
cyColaUrl	https://cdn.jsdelivr.net/npm/cytoscape-cola@2.1.0/cytoscape-cola.min.js
cytoscapeUrl	https://cdnjs.cloudflare.com/ajax/libs/cytoscape/3.2.8/cytoscape.min.js
dagreUrl	https://cdn.rawgit.com/cpettitt/dagre/v0.7.4/dist/dagre.min.js
cyDagreUrl	https://cdn.rawgit.com/cytoscape/cytoscape.js-dagre/1.5.0/cytoscape-dagre.js
weaverUrl	https://cdn.rawgit.com/maxkfranz/weaver/v1.2.0/dist/weaver.min.js
cySpreadUrl	https://cdn.rawgit.com/cytoscape/cytoscape.js-spread/1.3.1/cytoscape-spread.js
cyCanvasUrl	https://unpkg.com/cytoscape-canvas/dist/cytoscape-canvas.js

Or you can inject a bundle with all dependencies minified:

Vairable	Default Url
graphBundleUrl	https://chartfactor.com/wp-content/lib/cf_graph_bundle.min.js