Saturday 1st June, 2019 - 4 min read
This post is a work in progress.
I’ve been designing dashboards recently at work. It’s easy to throw a line graph, pie chart or other visualisation onto the page because it looks nice. But which visualisation should be used when? That’s what this article aims to answer.
What’s the best visualisation to use? The one that helps the user achieve the tasks. Understanding the users is key.
Stephen Few distinguishes dashboards from faceted analytical displays; a dashboard is a visual display of the most important information on a single screen to be monitored at a glance, while a faceted analytical display is a set of interactive charts providing multiple views on a common dataset that resides on a single screen and is used to analyse the information .
Decide whether your user needs to monitor information at a glance (the shift leader at an airport monitoring the average time for passengers to pass through security) versus analyse the information in depth (a market researcher analysing how sales for multiple products vary across regions, seasons, etc.)
I’ve found that often both of these views are required. The user first wants a high-level summary of the information most important to them, but then they want to click into the data to understand why there is a change and what they should do about it. Therefore dashboards click through to deep analytical displays.
What’s the best visualisation to use? The one that helps the user achieve their tasks. You need to understand the questions the user is asking of the dashboard to build the dashboard effectively.
For example, the key question for a dashboard to monitor the launch of a new product will probably be: Was the product launch successful?
You need to dig a stage deeper and understand the sub-questions that will help answer this primary question. Questions your dashboard will need to answer may include:
- How many users have used it?
- What is the user growth over time?
- How many errors have occurred?
- What are the most used features?
- What are the least used features?
- Has support ticket volume changed?
For each question, try to determine:
- The importance of this factor
- How frequently this will be asked/needed
- The depth needed for follow up
I find this helps determine where the visualisation should be displayed (key questions should be in the top left) and the volume and fidelity of information to display.
There are many visualisations available:
- Line chart
- Bar chart
- Pie chart
Here’s one framework for determining which visualisation to use.
Use a line chart to visualise a change in number over time. E.g., user growth over the last 6 months.
Use a bar chart. Used for displaying ordering/ranking. E.g., top used features.
Use a column chart. Displays the distribution of categories if the categories add up to 100%. Shows the part to whole relationship.
Use a line or column histogram. Helps to understand distribution across a population. E.g. shoe size, frequency of visits per user.
Use a scatter chart for two variables and a bubble chart for three variables. Measures correlation between factors.
Mark aside: No pie charts - why pie charts should rarely be used.
Some cheat sheets I’ve found from other articles online.
Enjoyed this post? ❤️
Subscribe by email or RSS.