- Introduction to D3
- Common Chart Types
- Base D3 Integration Pattern
- Binding Events to D3 Controlled Elements
D3.js is a JavaScript library for manipulating documents based on data. D3 helps you bring data to life using HTML, SVG and CSS. D3’s emphasis on web standards gives you the full capabilities of modern browsers without tying yourself to a proprietary framework, combining powerful visualization components and a data-driven approach to DOM manipulation.
As D3 is a well established library, we won't go very deep into D3 specifics in this article, but rather focus on how to properly utilize D3 in Meteor applications with full reactivity.
The good news is that D3 integrates very well with Meteor applications to create collection driven visualizations. Only minor changes are needed to adapt most D3 examples to be fully reactive to any data changes. Since D3 directly manipulates the DOM to create the data visualization, the Meteor rendering engine (Blaze) is not very involved in the actual rendering as D3 handles the creation and manipulation of SVG elements which make up the visualization. Meteor covers everything but the rendering as it manages the backing data, orchestrates the reactive updates, and binds events to the visualization elements.
The web version of Interactive Data Visualization for the Web is an excellent and in depth resource covering the most common use cases of D3. (Several of the example graphs were adapted from this source)
Since D3 can be used to produce wide array of data visualizations, a good way to start on building your own is to look over the many high quality examples and find one that is close to what you are looking for. A multitude of excellent example graphs can be found in the following galleries: https://github.com/mbostock/d3/wiki/Gallery, http://bl.ocks.org/mbostock
The D3 API Reference can be a useful resource when working with D3.
In the demo application for this article we have provided example code for most basic chart types common in web applications. Each of these charts is fully reactive to data changes in documents, adding or removing documents, and sorting.
Pie Chart - Source
Bar Chart - Source
Scatter Plot - Source
Line Chart with Time Data - Source
We will use the bar chart example to talk about the basic pattern used to integrate D3 into a Meteor application. This same basic pattern was used for all common chart examples and should be extendable to any D3 visualization within a Meteor application.
We start by defining an empty SVG element in template markup
<svg id="barChart"></svg>
Not much here at all, this is really just a placeholder for D3 to do its work.
All D3 code goes in the rendered callback for the template. Some code has been omitted using ... for brevity and to more clearly illustrate the pattern.
Template.barChart.rendered = function(){
//define constants, height/width
...
//define scales and axises
...
//define key function to bind elements to documents
...
//define the SVG element by selecting the SVG via it's id attribute
var svg = d3.select("#barChart")
...
//declare a Deps.autorun block
Deps.autorun(function(){
//perform a reactive query on the collection to get an array
var dataset = Bars.find(...).fetch();
//update scale domains and axises
...
//select elements that correspond to documents
var bars = svg.selectAll("rect")
.data(dataset, key); //bind dataset to objects using key function
//handle new documents via enter()
bars.enter()
.append("rect")
...
//handle updates to documents via transition()
bars.transition()
...
//handle removed documents via exit()
bars.exit()
...
.remove();
});
};
Following this basic pattern allows for relatively straightforward refactoring of most D3 examples to be driven via Meteor collections. The key part is separating the D3 code into what gets run once to initiate the visual (before the Deps.autorun block) and what gets run on each collection query result change (inside the Deps.autorun block).
It is also possible to use observe instead of Deps.autorun, see alternative options.
We want our visualization to update reactively whenever there is a change in the results of the collection query, contained in the variable dataset.
var dataset = Bars.find(...).fetch();
Since we encased all of the D3 code that updates the visualization within the Deps.autorun block, we already have reactivity when the results of the query change in any way. All we have to do is make sure that the D3 visualization update code handles added, changed, and removed documents. We use D3 to accomplish this by selecting a group of elements from the SVG and binding them to a data array.
var bars = svg.selectAll("rect").data(dataset, key);
D3 compares the supplied data to the existing graph and computing the needed changes. Any elements that are to be newly created are returned by enter() while any elements that should be removed are returned by exit(). By using these two D3 functions coupled with an update to any documents that may have been altered, via transition(), we have all possible data changes accounted for.
Once we have the reactivity supplied by Meteor and the handling of data changes by D3, we can make simple changes to the collection documents and the graph will automatically update itself. Since the visualization code is reactively driven by the collection documents, the source of the data change is not relevant. The changes can be made via event handlers, as shown below, or by any other mechanism.
//basic insert/update/remove operations on the collection cause a reactive update to the visualization
Template.barChart.events({
'click #add':function(){
Bars.insert({
value:Math.floor(Math.random() * 25)
});
},
'click #remove':function(){
var toRemove = Random.choice(Bars.find().fetch());
Bars.remove({_id:toRemove._id});
},
'click #randomize':function(){
//loop through bars
Bars.find({}).forEach(function(bar){
//update the value of the bar
Bars.update({_id:bar._id},{$set:{value:Math.floor(Math.random() * 25)}});
});
}
});
For best performance, we avoid running the collection query once to init the graph and then once for each change. Light refactoring may be required to adapt some D3 examples to have just one reactive query inside the Deps.autorun block.
To ensure that each collection document is correctly bound to it corresponding graph element(s), you'll need to use a key function.
//in rendered callback
var key = function(d) {
return d._id;
};
This function is used to specify how data is joined to elements. Since our graph is collection backed, the _id property makes sense, but you could use any other property that is unique to the document.
We then use this key function as part of our select in the Deps.autoun portion of the rendered callback.
//in Deps.autorun
var bars = svg.selectAll("rect").data(dataset, key);
Now D3 will use the _id property to link each bar to a single document. See the API reference on this topic for more details. Without this key function, the graph will still work, but certain changes to the collection query results will produce multiple visual effects in the graph. Removing a document is especially impactful as the index of all preceding elements is adjusted and those elements are redrawn/transitioned.
Since we are passing the result of a collection query to D3, we have full control over the order of the data that D3 will use to render. For an example like the bar chart, the collection query can be modified to sort by the value property of the document, which is the height of the bar. We supply a sort specifier to the collection query to return results in ascending or descending order which will reactively reorder the elements within the visualization.
var dataset = Bars.find({},{sort:{value:1}}).fetch();
This is a simplified example for illustrative purposes. The bar chart example provides a more complex 3 state (none, asc, desc) sorting toggle using the session to trigger a reactive update of the query results. You can use a dynamic toggle like this or a sort specifier could be hard coded into the query to always provide consistently ordered results.
We can use native Meteor event maps just like our normal template markup to bind to elements controlled and created by D3. This allows the visualization to be more tightly integrated with the rest of your application as events on your D3 created elements are just like events on your normal template created elements.
We'll need some context about what was clicked in the event handler, so we start with setting the data-id attribute of the rectangle to be the _id property of the document when the rectangle enters the graph via an attr() method call.
//in Deps.autorun block
bars.enter()
.append("rect")
... //other attr calls omitted for brevity
.attr("data-id", function(d){
return d._id;
});
Then in the template event map, we can pull the _id from the event target.
Template.barChart.events({
'click rect':function(event, template){
alert('you clicked a bar for document with _id=' + $(event.currentTarget).data("id"));
}
});
Now that we know what was clicked, we can act on this event. This simple example just does an alert, but this pattern could easily be extended to provide a drill down behavior or display a secondary detailed view of a given document.
The previously mentioned options are the recommended implementation for all meteor applications. However, the following options are technically valid and could be useful or desired in certain applications.
See the source of the alternative version of the scatter plot for a full example
var Observe; //placeholder for observe, needed to stop it on template destruction ... Template.scatterPlotObserve.destroyed = function(){ if(Observe) Observe.stop(); //stop the query when the template is destroyed }; Template.scatterPlotObserve.rendered = function(){ //define constants, height/width ... //define scales and axises ... //define the SVG element by selecting the SVG via it's id attribute var svg = d3.select("#scatterPlotObserve") ... //declare a transistion function to animate position and scale changes var transition = function(){ //find documents needed to scale domains ... //update scale domains and axises ... //handle transistions to documents via transistion() ... }; //transistion once right away to set the axis scale transition(); //decalare a debounced version, only runs once per batch of calls var lazyTransition = _.debounce(transition, 50); //declare an observe on a query Observe = Points.find({},{fields:{x:1,y:1}}).observe({ added: function (document) { //handle new documents svg .append("circle") .datum(document)//attach data .attr("id", "point-"+document._id);//set an id ... lazyTransition();//transition }, changed: function (newDocument, oldDocument) { //handle updates to documents svg .select("#point-"+newDocument._id) .datum(newDocument);//update data lazyTransition();//transition }, removed: function (oldDocument) { //handle removed documents svg .select("#point-"+oldDocument._id) .remove();//remove element lazyTransition();//transition } }); };Important differences from the
Deps.autorunpattern
- No use of D3's
enter()orexit()- Meteor is tracking the changes instead of D3- Extra care must be taken to stop the observe and to manage transition calls
- No attachment of a full dataset to a set of objects, just attach/update data for each object individually
It is possible that this approach is more performant than the preferred and more simplistic
Deps.autorunpattern. However it is significantly more complex and example D3 code requires more refactoring, thus theDeps.autorunapproach is recommended as the preferable option.
- could we fork d3 to be a component? let blaze do the rendering?
- line chart example could use real data - like https://coinbase.com/api/doc/1.0/prices/historical.html