Build a Scatter Plot with D3 v4

Share this video with your friends

Social Share Links

Send Tweet
Published 7 years ago
Updated 5 years ago

Scatter plots, sometimes also known as bubble charts, are another common type of visualization. They’re extremely versatile thanks to their ability to display multiple dimensions of data simultaneously using x and y position, opacity, color, and even shape. This lesson introduces the SVG circle element as part of building a robust scatter plot.

[00:00] As our starting point here, we have a nice blank chart with margins and axes set up, both of the axes running from zero to 100 with a simple linear scale. I've pulled in some sample data for this example.

[00:13] We've got an array of objects, and each object has a country name and code as well as the population, life expectancy, and average health care cost for that nation. We're going to plot these items on our chart using the population, expectancy, and cost properties.

[00:31] We come back over here to our code. The first thing we need to do is load in our sample data. We'll do that using the D3.json method and we'll tell it to load in our data.json file, which is in the same directory that we're working in right now.

[00:43] Go down here and grab our scales, so we can put them inside of our data loading callback. For our Y scale, we're actually going to use the life expectancy property here. What we can do is come back here and instead of just having the hard coded domain of zero to 100, let's get the extent of our data's expectancy property.

[01:04] We'll say D3.extent, pass in our data, and then the accessor function is going to return that expectancy property. We'll copy this and move down here to our X scale and we're going to say D.cost in this one, because we're going to use the health care cost as our X axis variable.

[01:23] If we save this we should get an update of our axes here, which we do. We've got zero to 10,000 on our X axis and 56 to 84 on our Y axis. You'll also notice that I've added these .nice method calls to our scales in order to make the axes end on more round numbers than they may otherwise, based on just the data.

[01:46] Now we've got our X scale and our Y scale set up. In this case for our scatter plot, we're going to be using circles. We actually need to create a new scale which we'll use for the radius of our circles. I'm just going to call this one "R scale."

[02:02] This time we're going to a square root scale, so D3.scale square root. We're using that because when you're using circles, in order to scale the area of the circle proportionally while setting the radius, you need to use the square root scale otherwise you get sort of out of whack proportions.

[02:24] In this case we're going to say our domain is going to start at zero and then we are going to use the D3.max method to get the maximum population, which is what we're going to use to set our radius on. The range of our R scale is just going to be zero to whatever we want the maximum radius of our circles to be. We'll say 40 right now.

[02:48] Now that we've got all of our scales set up, we can actually go about creating our circles. If we say SVG.selectall circle, so SVG circle elements is what we're going to use here. We're going to set our data here and do our data join. Let's move this down.

[03:07] Within our enter selection, we want to append a new circle. SVG circle elements have specialized properties which are CX and CY rather than just regular X and Y. That stands for Center X and Center Y, so it's drawn using the center point and then grows out from there.

[03:27] Remember, we set our X scale to use the cost property and the Y scale to use the expectancy property, so those are the values that will pass to our scales in order to set these attributes.

[03:39] The last thing we need to set the R attribute for the radius, and for that we will use our new R scale. We'll say R scale, D.population. Let's give it a fill style so we can make sure we see it. We'll use steel blue again.

[03:55] Save that and it doesn't render. Ah, that's because our R scale here, we forgot to pass in the data. If we save that, now we get a nice scatter plot with some big circles, some small circles. Let's actually add a little bit of opacity to these circles so that we can see where they overlap. There we go, now we can see where the circles overlap each other.

[04:19] That's a basic scatter plot. We have now communicated three dimensions of data here. We've got the cost along the X axis, the life expectancy along the Y axis, and the population of the country based on the size of the circle.

[04:36] It's a little hard to tell what is what around here because there are no labels. To do that, we could do a couple of things. One option would be to plot a bunch of text items in the same position as the circles, and lay them over top.

[04:55] If you want to animate things, that can get a little tricky and things like roll overs. It's actually better if you group them with an SVG graphics container. Let's see how we do that.

[05:06] The first thing we're going to do is change it so we're not creating circles here, we're creating graphics containers. We don't really want to do this and say that we want it to select all of the graphics containers because our axes and other things on the chart potentially are graphics containers, and we don't want them selected.

[05:25] We're actually going to use a CSS class as our selector criteria here. We're going to call it ball. In order to make sure that we have the appropriate CSS class applied, when we create our graphics down here, we'll go ahead and set class to ball here.

[05:42] Let's give ourselves a little bit of space here and we're going to actually assign this to a variable called circles. What we're doing here is we're creating a bunch of graphics containers with the CSS class of ball and we're assigning that selection to a variable named circles.

[06:04] In order to position these graphics containers where we want them, we need to use the transform attribute. The transform attribute is going to need to call the translate function. We're actually going to say D, pass that in, and we're going to return a call to translate.

[06:23] We're going to grab these values from down here. The first value that gets passed to translate is our X scale calculated value, the second one is our Y scale. Now we've got those there. We close this off.

[06:39] Now if we say circles.append circle, so remember, circles is our selection of G elements. We're going to append a new circle to each one of them. Since we set the position on the G element itself, we can actually set the CX and CY to zero.

[07:00] We'll still set our radius here using our R scale, but if we save this we should have the same exact output, but we don't. Something is wrong here. That is because I misspelled transform. Let's fix that and there we go.

[07:15] We're back to where we were before, but now we actually have circle elements within graphics containers. Now that we have that in place, we can go ahead and create our labels. We'll again say circles.append, and this time we'll append a text element.

[07:33] We're going to set a style on these called text anchor and we're going to set that to middle because we want the text to be center aligned. Let's fix that.

[07:44] Then if we set the fill style so that we can actually see these things -- we'll set them to black for now. The last thing we need to do is actually set the text itself. We use another text here and we'll say text is D.code, and so now we have our labels here in the middle of our circles.

[08:05] You can see they're a little bit off. We can actually set the Y attribute here, maybe like four pixels down. There we go.

[08:15] Now if we actually go in and inspect our markup here, we can see that we have apparent graphics container here. We've got a bunch of graphics containers. They've each got the ball CSS class. If look inside there, we've see our circle and our text element, so everything is grouped together nicely, easy to animate, easy to interact with.

[08:37] There you go. You've got a labeled scatter plot all based on real data.

Nenad
Nenad
~ 2 years ago

A suggestion is to clip the circles which are overlapping with yAxis and show the approach how to clip them.

Markdown supported.
Become a member to join the discussionEnroll Today