Search
Close this search box.

The Point-as-Line Chart, PAW and Beyond – I Was Wrong

I have a confession.

In my article about the point chart, I disavowed IBM’s suggestion that a point chart is a “line chart without the lines”.

True, this is neither the original purpose of what was originally known as the Cleveland dot plot nor the extent of what it can do in Planning Analytics Workspace (PAW). However, it’s still a suitable use of the point chart. The user connects the dots, and they see the lines.

Don’t get me wrong, your audience won’t connect any dots with this:

A point chart showing the revenue for five different product lines for 2012, 2013 and 2014. Each year has a vertical line, and each product line has a point on a each line. Higher revenue products lines are higher up on the line. Click to visit the IBM PAW documentation on the point chart.
The point chart in IBM’s documentation. Click for source.

However, position your data right, and your viewers can follow its journey.

Rationale

Why use the point chart over a connected line chart?

There are many ways to get from point A to point B. Every single line you draw between points is an assumption. Straight lines? Curved lines? What type of curve? Does it have to be smooth? Must it go through every single point? With the point-as-line chart, there are none of these assumptions. You don’t get the connectivity of a line, but you don’t get the risk, inaccuracy or clutter of an unnecessary line either. Your data can speak for itself.

A series of charts. The first has a series of points with no lines between them. The other charts each connect the points with a different method of drawing lines (basis, basis open, cardinal, cardinal open linear, monotone, step, step before and step after).
The PAW line chart has nine interpolation methods, some more audacious than others

How the Point-as-Line Chart Works

Even without these connectors, point charts can still work as line charts. This chart from TIME demonstrates it well.

A point chart which shows primary school completion rates for male and female students from 1974 to 2018. As time progresses, from top to bottom, the gap between male and female students closes, and completion rates for both groups improve.

This chart shows global primary school completion rates for male and female students. The chart uses three Gestalt principles to group them together:

  • Proximity: the points for the two series are close to one another.
  • Similarity: each series of points has its own colour.
  • Continuity: the value of the points doesn’t change much from one year to the next, so the points flow into one another.
A diagram of eight Gestalt principles.
Two rectangular groups of six circles represent the proximity principle - the close shapes are visually grouped together.

Outlined and filled circles mixed together represent the similarity principle - the viewer groups the outlined and filled shapes together, even when they're not next to each other.

A group of circles with a filled box behind two circles and an outlined box behind three other circles represents the enclosure principle. The boxes group the circles inside together.

A grid of twelve circles with a line through four of them represents the connection principle. The line groups the four points together.

For the continuity principle, two paths of circles cross. They are filled before they meet, and outlined after they meet. Because they're arranged in continuous paths, the viewer groups the points on the same path together.

For the closure principle, circles form the outline of a box. Because they're arranged in a closed shape, the viewer perceives the shape.

For the figure-ground principle, a path of dots leaps out from the page left-to-right, with a wavy grey area behind it. The user perceives the figure (circles) and ground (grey area) separately.

For the common fate principle, a line of circles has arrows pointing out of it, with the direction alternating. The viewer perceives points going in the same direction as part of the same group.
Above: multiple ways to group visual objects together. Click image for source.

Put connection against any of these principles on their own, and connection would be stronger.

Three sets of points. The left set of points is spread out, but lines between them create a visual link. The middle set of points is different colours, but the lines between them create a visual links. The third set of points form a flowing path, but due to lines between the points from one path to the other, the viewer still groups the points by the lines instead of by the paths.

But when you combine the three principles, there’s a clear path to follow. The link across years within the same gender is stronger than the actual lines on the chart.

Two series of points. Each has its own colour, and the points on each path are close together and form a smooth path. Even though there are lines connecting points from one path to the other, the viewer perceives the paths before the connected pairs of points.

Compare the TIME chart to the chart from IBM’s documentation. The only principle linking the points by product line is similarity. The years are too few and far apart for proximity or continuity to group the points. Thus, the connectivity from the grid lines wins out, and the viewer groups the points by year.

A point chart which shows primary school completion rates for male and female students from 1974 to 2018. As time progresses, from top to bottom, the gap between male and female students closes, and completion rates for both groups improve.
A point chart showing the revenue for five different product lines for 2012, 2013 and 2014. Each year has a vertical line, and each product line has a point on a each line. Higher revenue products lines are higher up on the line. Click to visit the IBM PAW documentation on the point chart.

When to Use the Point-as-Line Chart

A template point-as-line chart. The horizontal axis is labelled "Dimension with several ordered elements (often time)", and the vertical axis is labelled "Ordered dimension (usually numeric)".

The point-as-line chart is but a tool for data-driven decision-making. Your data is more important. Before you make your point-as-line chart, check your data is suitable:

You have two ordered dimensions: these are just dimensions where the elements have a clear order to them – 2024 comes after 2023, July comes after June. The order is usually low-to-high, but it can also be cyclic.

One of these ordered dimensions goes along your item axis – it’s usually time-based. Your item axis dimension should be discrete, not continuous – you should be able to split it up into individual elements which can’t be split any further. 2023.5: it’s a number, but it’s not a year. If your item axis is a quantity or otherwise continuous, then consider a scatter chart instead.

The other ordered dimension goes along your value axis. As long as it has an order, it will work – continuous or discrete.

You have enough detail: how many elements are in your item dimension? The more you have, the easier it is for a viewer to follow the data. Consider a line chart if you have too few points. Column charts are also a suitable alternative, especially if your elements are higher-level consolidations with large fluctuations.

Four point-as-line charts. The charts are the same size, but each chart has more data than the next. The points are closer together, so the viewer can mentally connect them more easily.

Your two dimensions speak for themselves: you can use colour to add a third dimension to your point chart. Your chart will have multiple lines, one in each colour. This adds information, but it also disturbs the point-as-line chart’s delicate balance. As each line’s points get close and intersect, the individual paths blur together. If you can guarantee your lines will stay clear of one another, then a point-as-line chart can work, such as the TIME example above. Otherwise, a line chart will display your series more clearly.

Three charts. The first has a single series of data in a point-as-line chart, and it is easy to see the series. The second chart is the same, but it has two more series of points, each in its own colour. The individual series are hard to distinguish. The third chart is the same as the second, but it is a line chart: lines in matching colours connect one point to the next. The individual series are easy to distinguish.

Your data follows smooth paths: proximity and continuity are key to an effective point-as-line chart. Without lines between your points, the viewer can connect the dots however they see fit. So, point-as-line charts work best with points that stay close to one another as the item axis moves from one end to the other. This way, the viewer won’t get lost. If you expect volatile data, consider a line chart.

Two point charts. In the first, the points follow a smooth path, and the viewer can easily see them as a unified series. The points in the second chart  an erratic path, so the viewer perceives a directionless cloud of points instead of a smooth line.

Line charts are also better if your data has outliers. Since outliers are so far from the rest of your data, viewers can easily miss them entirely in a point-as-line chart. This is excellent if you want to mislead your audience and sweep your outliers under the rug, but not so good if you want to make robust, well-informed decisions with your data. Or if you want to sleep at night.

Two charts. Both have the same data: several high points, and one low outlier point in the middle. The first chart has no lines between points, and the viewer can easily miss the outlier. The second chart has lines between adjacent points. Here, the outlier is very hard to miss.

How to Use the Point-as-Line Chart

PAW makes it easy to create a point-as-line chart.

Step By Step

A template point as line chart. The horizontal axis is labelled "Item axis", the vertical axis is labelled "Value axis", and the legend is labelled "Colour".

To make a point-as-line chart in PAW, you can either start with a point chart or an exploration.

Point Chart Method

This is the easier and more intuitive method.

  1. Open a PAW book and go to the “Visualizations” tab on the left-hand side of the screen.
  2. Click on “Point”. An empty chart should appear in your book.
  3. Go to the “Data” tab on the left-hand side of the screen, and drag and drop your chosen cube or view onto the empty point chart.
  4. Select the chart and open “Fields” in the top-right corner. Move the dimensions and change the subsets as desired.

Exploration Method

The exploration method is best if you have custom MDX you want to deploy. This is more complex than the drag-and-drop method, but it allows you to create more nuanced charts.

  1. Create an exploration: this can come from the “Data” tab or the “Visualizations” tab.
  2. Arrange your dimensions:
    • Item axis: First dimension on rows
    • Colour: second element on rows
    • Value axis: Columns
  3. Where it says “Exploration” at the top of the page, click and select “Point”.

Either way, once you’ve made your chart, open “Properties” in the top-right corner to make visual changes.

Tips and Tricks

For your point-as-line chart to really work, keep your Gestalt principles in mind. There’s no connectivity here, so proximity, similarity and continuity are the principles at work. They’ll need some help to keep your lines together. So, keep the following in mind:

Minimise the number of series: this strengthens the continuity of each line. With less clutter, your viewer can focus more easily on each line. Consider consolidating the dimension in the colour field here, if you use it at all: the “repeat” fields are a suitable alternative.

Five charts. The first has a single series of data in a point-as-line chart, and it is easy to see the series. The second chart is the same, but it has two more series of points, each in its own colour. The individual series are hard to distinguish. The last three charts are the individual series from the second chart, one series per chart. The series are easy to distinguish in their own separate charts.

Squish it in: to maximise proximity, bring your points close together. You can do this by adding more elements, but you can also make your chart smaller in both width and height. There’s no objective way to choose your chart’s aspect ratio, but be careful not to exaggerate or flatten any trends in your data.

The same point as line chart, in three different sizes. When the chart is large, it is harder for the viewer to mentally connect the points. When the chart is small and the same aspect ratio, the viewer can easily connect the points. However, when the chart is small but much skinnier, the viewer can connect the data, but the trend in the data is exaggerated.
Rescale at your own risk.

Point shape: PAW offers you many point shapes. It won’t be hard to choose between them, since almost all of them are terrible at their job. Let me explain:

The distance between the centres of adjacent points will always be the same, at least horizontally. However, what we perceive is the distance between the edges of the points.

Two sets of squares. The first set has no lines between them, and the viewer mentally measures the distance between them from one edge to the other. The second set shows the centre-to-centre and edge-to-edge distance between the squares.
What you see is not what your chart sees.

Even for the same shape, the edge-to-edge distance can change as the angle between two points changes. Due to the proximity Gestalt principle, the audience will perceive a stronger or weaker trend there, not because there is one, but because the apparent distance between the points is distorted.

Circular and square points arranged in horizontal and diagonal lines. The circles are the same distance either way, but the squares appear closer arranged diagonally than they do horizontally.
Circles will always appear the same distance from each other regardless of angle, but other shapes will distort distances.

So, to minimise this distortion to your series, each point on the perimeter of your shape should be the same distance from its centre – this is a circle by definition.

As well as the typical circle, PAW offers the ‘donut’ and the ‘donut with center’.

A circle, a hollow circle and a hollow circle with a solid circle inside.

The donut shapes are less visually prominent than the circle, and they work especially well for dense, overlapping data. If you’re using a donut shape, then make sure your colours contrast especially well with the background and each other. Speaking of which:

High contrast colours: your viewers need to distinguish each series from one another. So, the colours in your palette should contrast well with each other and the background, especially in luminance (light-dark). All other colour tips from the original point chart article also apply.

The Journey

Despite previous convictions, point charts actually work as “line charts without the lines”. However, be sure to chart your journey well. Without any lines between you’re points, you’re free from paths trodden before, but you can also get lost. Ensure you have a small amount of smooth data, and format it so your audience can easily get from one point to the next. This way, your audience will arrive at their destination: your data, ready to inform your decisions.

Need Help with Point-as-Line Charts?

Please reach out if you need any assistance with PAW visualisations – we’d love to help!

  • This field is for validation purposes and should be left unchanged.

Leave a Reply

Your email address will not be published. Required fields are marked *

Log In