• Home
  • Home

Dot charts

home/Documentation/Dot charts
Expand All Collapse All
  • How to create a scatterplot
  • How to create a bubble chart

How to create a scatterplot

247 views 0

Content:

  • When to use a scatterplot
  • Negative aspects
  • How to create a scatterplot on BStreams
  • Example
  • Common mistakes

The scatterplot shows the correlation between two numerical variables, represented along both the X and Y-axis. While in most charts, the dimension is plotted on the axes, in the scatter chart, the points within the chart represent it. Thus, when a third optional metric occurs, its value reflects the dots’ size.

This kind of chart is usually used in the scientific field, and it’s because of this, it is perceived as challenging to understand. In contrast, in the marketing field, it’s commonly used to represent customer clusterization.

When to use a scatterplot

Scatterplot is a powerful data visualization commonly used to display the relationship or correlation, both negative or positive, between values and to find anomalies in datasets. Finally, the third optional metric is helpful to simplify the identification of massive countries, customers, and quantities, just to name a few.

In short, want to determine whether there’s a potential relationship between the two values? A scatterplot can help you.

Negative aspects

Scatterplot is Scatterplot may be a challenging data visualization to understand for a novice audience because of its metrics on both the X and Y-axis. Moreover, adding the third metric might increase the complexity of interpretation.

To avoid this, remember to include descriptive labels to make the data in the scatterplot easier to understand.

How to create a scatterplot on BStreams

The scatter plot on the BStreams is designed in such a way that dots represent dimension and the X and Y-axis correspond to metrics. When you add the third metric to the chart, it manages the diameter of the dots. Creating a scatterplot on BStreams implies that you always have to put in your dataset two or more numerical variables and a qualitative one.

Here are the steps to create a scatterplot on Bstreams:

  • Add a new project
  • Drag the scatterplot and drop it on the canvas
  • Select the sample dataset
  • Choose one dimension and two metrics (mandatory)
  • Add an optional third metric to generate the dot size
  • Click apply to show the result

Example

Look at the following example showing the correlation between the total duration (X-axis) and the total budget (Y-axis) of the 2020 highest-grossing films. The third metric corresponds to the total gross, and it sets the dot’s size. The highlighted dots have been colored, choosing the highest gross per film with a value >2000 mln dollars.

Common mistakes

Overplotting: it is probably the most common mistake when talking about scatterplots. This mistake occurs especially when the dataset is significant, and dots overlap, making the chart look confusing. The simplest way to fix this problem is to reduce the dots’ size.

Data is not related: scatterplot is not the best option to adopt when there’s no correlation at all with variables. For example, suppose you want to survey the Los Angeles Lakers players and want to put together two measures like the player’s height and the times in a week they dine out. In that case, it’s pretty evident from the start; that there’s no correlation between these two variables.

Confusing correlation by causation: when we visualize two numerical metrics in a scatterplot, it doesn’t necessarily imply that changes in one metric are responsible for changes in the other.
This is particularly true referring to the common thesis according to which “correlation does not imply causation” which highlights the impossibility of “deducing a cause-and-effect relationship between two events or variables solely on the basis of an observed association or correlation between them” (Wikipedia).
There is a high possibility that the given relationship may be driven by a third measure that affects both the metrics, that the causal link is inverted, or that the pattern is just coincidental.

Want to learn more about scatterplots? Email us at support@bstreams.io

If you haven’t checked our blog yet, give it a look and let us know what you think about it.
Read more here

Was this helpful?

Yes  No
Related Articles
  • How to create a bubble chart
  • Terms
  • Privacy
  • Cookies
  • Copyright © 2022 BStreams. All rights reserved.