How to Adapt a D3.js Template in a Web App

D3.js is a state-of-the-art library for data visualization. Check out the D3 gallery for stunning and beautiful examples. Happily, many of these visualizations include their source code, so that you can easily duplicate them.

For example, the parallel coordinates chart, created by Mike Bostock, is given with the generating D3 code and data! This is a cool and useful data viz that allows you to quickly visualize a multi-dimensional (but relatively small) dataset. You can immediately spot correlations across dimensions and uncover clusters. In this interactive visualization, you can explore the data in depth by filtering values on each dimension with a brush tool.

This brief tutorial replicates the parallel coordinates chart in a Dataiku DSS web app. The final version of that web app can be found in a project on the Dataiku gallery.

"Example of a parallel coordinates chart"

Prerequisites

  • The tutorial on the basics of standard web apps is suggested, but not required.

Upload the data and create a new web app

The parallel coordinates chart is illustrated on a dataset of car specifications. In your project, create the cars dataset by uploading this CSV file.

From the Code menu, create a new “standard”, empty web app. You can delete all of the sample code.

To access the data in your web app, open the Settings tab and click on “Configure” in the Security section. In the dataset list, find the cars dataset and allow your web app to read it. Then, import the D3.js library on the main page of the Settings tab. As a reminder, the “Basics of HTML/JavaScript Web Apps” lesson in this course covers this process in greater detail.

Understand the overall code structure

Many D3 code samples, given in the gallery or bl.ocks.org, have the same overall structure.

<!DOCTYPE html>
<meta charset="utf-8">
<style>
    /* CSS  code */
</style>
<body>
    <!-- HTML code -->
    <script src="https://d3js.org/d3.v3.min.js"></script>
    <script>
        // JS code
    </script>
<body>

To replicate the D3 visualizations in your web app, you will simply need to copy the CSS and HTML code in the corresponding panels of the web app editor. For the JavaScript code, it requires a little more work, as we will see promptly.

Copying the HTML code

In the parallel coordinates chart example, there is no HTML code written within the <body> tags, and thus the HTML panel of your web app code editor should be blank. Be sure to remove any sample code.

Copying the CSS code

Here is the CSS code, defined within the <style> tags, that you should copy into the CSS panel of your editor.

svg {
  font: 10px sans-serif;
}

.background path {
  fill: none;
  stroke: #ccc;
  stroke-opacity: .4;
  shape-rendering: crispEdges;
}

.foreground path {
  fill: none;
  stroke: steelblue;
  stroke-opacity: .7;
}

.brush .extent {
  fill-opacity: .3;
  stroke: #fff;
  shape-rendering: crispEdges;
}

.axis line,
.axis path {
  fill: none;
  stroke: #000;
  shape-rendering: crispEdges;
}

.axis text {
  text-shadow: 0 1px 0 #fff;
  cursor: move;
}

Adapting the JS code

The trickiest part in adapting a D3 template is always to shape the data in the format required by the data viz. In the parallel coordinates charts, the data in the D3 code is represented as the cars JSON array.

Generally, however, your source data is not in JSON format. In many D3 templates, the data is given as a CSV file, which is converted to JSON.

In the original D3 code, the data is thus read from the cars.csv file:

// D3 code
d3.csv("cars.csv", function(error, cars) {
    // D3 code
});
// D3 code

Then the D3 code defined inside the d3.csv() function is applied on the cars JSON array.

In our DSS web app, you will have to connect to your dataset (which can be stored in a great variety of formats and database systems) through the Dataiku JavaScript API.

In order to do this, we need to slightly modify the JS code. Without touching the body of the function, replace the original d3.csv() function name and parameters with the code below:

// same D3 code
function parallelCoordinatesChart(cars) {
    // same D3 code
}
// same D3 code

In other words, keep the entire D3 code unchanged, except for the call to the d3.csv() function, which is replaced by defining the parallelCoordinatesChart() function, which takes the cars JSON array as input.

Note

Be sure to remove the parenthesis and semicolon — ); — that were at the end of the d3.csv() function; they are not needed (and the parenthesis will indeed cause an error).

Now, we only need to connect to the cars dataset through the Dataiku JS API, in order to create the corresponding cars JSON array. Notice that, when you gave permission for your web app to read the cars dataset, an option to add a snippet calling to the dataiku.fetch() function could have been added.

You finally need to copy the JS code defined below into the dataiku.fetch() function. This code creates the cars JSON array and calls the parallelCoordinatesChart() function to create the chart.

dataiku.fetch('cars', function(dataFrame) {
    var columnNames = dataFrame.getColumnNames();
    function formatData(row) {
        var out = {};
        columnNames.forEach(function (col) {
            out[col]= col==='name' ? row[col] : +row[col];
        });
        return out;
    }
  var cars = dataFrame.mapRecords(formatData);
  parallelCoordinatesChart(cars);
});

That’s it, you have a running D3 data viz in your web app!

../../../_images/d3-webapp-code.png

Troubleshooting

If you’re having trouble, be sure you have carefully followed all the steps. The best way to debug is to use the JS console in your browser with the web app editor open.

You can also find a completed version of the web app in the Dataiku gallery.