Data visualisation allows you to see your data from a new perspective and helps you create improved overview. When you choose the right charts for the right purposes, you are able to compose a visually convincing and coherent story.
We will provide you with input for some of the most important considerations you need to make when working with data visualisation. In addition to this, we introduce you to some of the most commonly used types of charts and suggest the type of data visualisation most suitable for the different charts.
Consider the user in the set-up
Who are you designing the data visualisation for, and where do they want the data to be visualised? Knowing your audience will help you determine the data you need.
By customising the contents to the individual user or user group, creating the proper balance between informative and inspiring data visualisation becomes much easier.
Some users – internally as well as externally – may be satisfied with a simple pie chart, whereas others may want visualisations in detail that allow them to dig deeper into the insights collected. Designing just one dashboard for the entire organisation is rarely the optimum solution.
Take into account that describing what you want to see may be difficult without having some specific options. So help the interested parties by illustrating different options. I will also make it easier for them to express their wishes for what they want to see as well as the level of detail required.
Determine the goals with data visualisation
In order to organise your data visualisation, it we recommend that you focus on making a logical story. Once defined, you can dive into the most significant insights. It is important to define a clear set of goals before building your reports and dashboards. Once you have decided which story to tell, it becomes much easier to choose the right type of data visualisation.
Compose each report or dashboard to hold the answer to this basic question – “Which steps should be made, having these insights?” The overall goal of your visualisations is thus always to prepare the users for taking action.
Design the data visualisation to match the frequency of the decision. Is the decision strategic, and therefore only needs to be answered once, for example by large investments? Or is it operational and needs to be answered several times a day? The third option is tactical decisions that may require a weekly or monthly review?
Build comparisons and context for the data visualisation
Consider whether it makes sense to add tangible comparisons in order to generate context when presenting your data.
To prepare for best action, it is an advantage if the user has a clear goal or a benchmark to compare to from a previous period.
Presenting two charts next to each other showing contrasting versions of the same information over a given time period, such as monthly sales data for 2017 and 2018, will for example emphasize strengths, weaknesses, trends and opportunities on which action can be taken.
Optimise color selection and ranging on charts
It is important to select the colors carefully when setting up charts. Most people associate green with positive trends, whereas red is associated with negative. Choosing other colors for indication of positive/negative may be confusing.
Do not use colors close to each other in the color spectrum, but at the same time make sure that visualisation is not distorted by some colors being much stronger than others for no reason.
When composing a dashboard or a report with more visualisations, the color palette should be considered across charts – should the palette be the same for all, or should some charts be further emphasised. Colors may have a great significance for what catches the eye, and it may be of advantage to use the same colors.
Ranging the data visualisation will also affect what attaches the highest importance and/or the procedure the user must follow to achieve the best interpretation of the data.
Fast tips for data visualisation
Before looking at the use of various chart types, we here present you with a few mnemonics for the data visualisation.
The overall recommendation is to:
• Use clear labels, describing notes etc. to guide the users
• Avoid distortion of data and let the axes on charts start with 0
• Discard 3D – it confuses and presents no advantages
• Be consistent when selecting chart types across visualisations that puts up to comparison
• Consider how the user sees the visualisations – pc, table, phone
• Keep visualisations as simple as possible to strengthen the message and simplify decoding
Choose the right charts for data visualisation
Here are some examples of the most common visualisation types for the different data visualisations.
a) Line chart and Area chart
When you wish to visualise a change over a given period, a simple line chart is highly efficient.
In addition to this, you can easily compare different variables over a given period. However, we do recommend that you do not have more than 5-7 variables in order to maintain the overview.
Like line charts, the area chart is suitable for visualising a trend over time. Given that the area chart has filled areas, it is not be recommended in situations with overlapping lines.
This chart is most suitable when stacking values makes sense, for example for estimation of values as a ratio of the total in order to enhance the cumulative values.
b) Bar chart and Stacked bar chart
The bar chart is used for comparing sets within various categories.
Color coding makes it possible to compare a number of data in a clear manner.
Like area charts, the bar chart can also be stacked in order to estimate parts of a unity. Be aware not to include too many columns, as this will affect the overview.
A horisontal bar chart is evident for comparing rankings, such as a top-five list. Make sure to maintain a sensible order of appearance – either listed by value or some other logic such as alphabet or time.
c) Pie chart and Donut chart
The pie chart may have a slightly bad reputation, but that does not change the fact that a pie chart can work very well as data visualisation when used correctly.
Pie charts are especially useful for demonstrating the proportional composition of a specific variable. They are thus suitable in situations where overall estimations are sufficient and when the elements total 100%. Another advantage is that most users know the pie chart and instinctively know how to read it.
However, it is important to remember that it may be difficult to get the real value from a pie chart, if too many variables are included, and/or if these are close to each other in size. In such cases, we recommend to attach labels indicating the value for each of the variables.
If a pie chart holds more than seven variables, you should consider changing it to for example a bar chart. It may make it harder to decode the ratio of the total, but seeing the variables in relation to each other will be easier.
d) Scatter plot and Bubble chart
A scatter plot shows the values of two variables along two axes. The pattern between the points visualises the correlation between them. One of the advantages of a scatter plot is that it can contain way more elements than the chart types mentioned above.
A scatter plot is excellent for emphasising correlations between dimensions. Conclusions are thus highly drawn from the grouping (or the lack of groupings) of the points rather than from the individual points.
For scatter plots it is, however, important to remember that it takes a certain amount of points to make sense. This also goes the other way around – the more points included, the longer it takes to check the visualisation before it can be interpreted. Therefore, if the goal is to achieve a quick insight, the scatter plot may not be the right choice.
The bubble chart is very similar to a scatter plot, but it is characterised in that it allows showing the variation between three data items. The size of the individual bubbles represents the third variable. The chart is quite complex and may not decode fast, but it opens up to other analytic possibilities.
e) Radar chart
The radar chart is visualisations compare more quantitative variables.
Each variable has an axis starting from the centre point. All axes are arranged radially, with equal space and the same scale between all axes. Each variable value is drawn along its individual axis and all variables in a data set and joined together to form a polygon.
Radar charts are useful for showing variables with corresponding values as well as for indication of whether there are deviations among each variable. Radar charts can also be used for showing which variables in a data set that score high or low.
However, more than two polygons in a radar chart may be difficult to decode.
Likewise, too many variables may create too many axes, which will increase the complexity. It is recommended to keep the charts simple and with a limited no. of variables.
f) Speedometers
Speedometers use pointers and colors to show data, and it is an example of a quickly interpreted data visualisation.
They are suitable for showing a single value/a single goal within a quantitative context, such as compared to previous period of time or a goal value. It provides an efficient indication of an immediate trend.
Speedometers are useful for KPI’s and individual data points. Having only one data point, they are not evident for comparing different variables and are thus not directly useful to act on alone. Also, speedometers take up quite a bit of space on the dashboard, which is why it in some cases would be better to choose another chart type that can summarise more KPI’s.
g) Figure chart
The figure chart is one of the simplest visualisations, providing an instant image of a certain KPI, such as total sales, ratio of delivery, total no. of visitors etc.
Such figures could be assigned a color code indicating how the value is positioned compared to the determined goal.
For figure charts as well as for speedometers apply that you should avoid too many on you dashboard.
Likewise, it is important to have a clear indication of time and a describing label.
