Analytics enables a high level of visual customization to dashboards. You should ensure that the dashboards you create:
-
Present information accurately
-
Are visually appealing and easy to read
-
Have a consistent look and feel
-
Follow your organization's brand guidelines and requirements
This document provides best practices and recommendations to consider when you are creating your dashboards.
Filters
-
To reduce clutter, Filter Display should be set to Popover unless there are three or fewer options and the filter is used very frequently.
-
Filters should be listed in a logical order (typically most frequently used first).
-
Most filters should be Checkbox control type except where there are over a dozen or so items, in which case Tag List should be used. The Advanced control type should be avoided if possible.
-
Filter options should be presented in a logical order (e.g. Small>Medium>Large>Extra Large).
-
Filter labels should be user-friendly without abbreviations or special characters - e.g. “Customer Last Name” rather than “CUST_L_N”.
-
Filters should be linked where appropriate, based on the data.
-
When filtering by dates or timestamps, the date version of the field should be used (not month, week, hour, etc), unless you have a specific need to use another version.
Layout
-
Data should be organized to flow logically and “tell a story” when possible. For example, if the data includes several steps of the custom journey, the dashboard should follow that flow.
-
Important questions should typically be presented first.
-
Charts should be thematically grouped when possible - e.g. product features, satisfaction, consumer habits, or customer service. Text tiles can be used as group headers.
-
The layout should be designed to display properly at 100% browser zoom.
Data Presentation
-
The Y axis of a chart should typically be adjusted to accommodate the range of possible values. For example 0% to 100% for a percentage value, or -100 to +100 for NPS.
-
Legends should be avoided when possible as they require the user to look in two places to understand the data. When used, the order of values in the legend should match the order in the chart.
-
Decimal places should be sensibly used. For example, it is not useful to know the average respondent age is 43.2465 years.
-
Excessively long answer options should be abbreviated.
-
With choice questions, if there is a logical order to the options (e.g. Small>Medium>Large>Extra Large), that order should be used in charts. For options without a logical order (e.g. favorite ice cream flavor) the options should be presented in descending order of popularity so the user can quickly identify the top and bottom values.
-
Word clouds should be avoided as the current version does not provide useful information.
-
Charts should be consistent within the dashboard - for example if the survey asks “How do you like our service?” and “How do you like our products?”, those questions should use the same chart type and settings.
-
Pie charts should not be used when there are more than 4-5 values to display, as they are too cluttered.
-
When displaying data (especially open ends) in a table, include in the table any other relevant information - for example, if the question is “Why did you give the NPS rating you did?” you should also include the NPS value for context.
-
Grid questions, and groups of questions structured similarly to a grid, can be displayed in the following ways:
1. Use multiple charts, each with a text tile positioned next to the chart:
2. Create a bar chart with all of the values, but hide the legend, and add an HTML table next to the chart containing the question labels. This option will require work to line up the rows properly and may look inconsistent compared to the rest of the dashboard:
The text tile next to the visualization above is formatted using the following HTML code.
<table style="float:right; width=100%">
<tr height="85px">
<td> </td>
</tr>
<tr height="85px">
<td><b><p align="right">Customer service via the 1-800 line</p></b></td>
</tr>
<tr height="85px">
<td><b><p align="right">Documents you received in the mail</p></b></td>
</tr>
<tr height="85px">
<td><b><p align="right">Emails you received</p></b></td>
</tr>
<tr height="85px">
<td><b><p align="right">Financial Professionals</p></b></td>
</tr>
<tr height="85px">
<td><b><p align="right">Mobile App</p></b></td>
</tr>
<tr height="85px">
<td><b><p align="right">Website</p></b></td>
</tr>
</table>
Look and Feel
-
Use your organization's brand color scheme for the dashboard, if possible.
-
For charts with a legend, it may not be viable to use a corporate color scheme if the palette is too limited - in that case other colors which are visually complementary should be used
-
Bright “warning” colors should be used sparingly or they will lose their impact - if half of a dashboard is bright red the user will not know where to look. They should also be used consistently. If the survey asks “Are you happy with our service?” and “Did you have any questions?”, the “No” answers should not both be formatted red, because they indicate different things.
-
You can use Markdown syntax to apply formatting to text tiles.
Standard Visualization Settings by Chart Type
Below are standard settings for some of the most common chart types to create a clean and consistent visual style. To access these settings, edit a tile and then click the gear menu to the right of the visualization section. Any settings not included here should be left as default.
Table
-
Plot
-
Table Theme = White
-
Show Row Numbers = No (unless needed)
-
-
Series
-
Truncate Text = Yes unless working with open-ends
-
Size Columns to Fit = Yes
-
Customizations > Label = rename to be user-friendly
-
Customizations > Column Width = size to fit content
-
Column / Bar
-
Series
-
When using Series Positioning=Stacked Percentage, use the Count field instead of the Percent field, so that a mouseover will display both the count and percent
-
-
Values
-
Value Labels = Yes
-
Font Size = between 12 and 20
-
-
X
-
Show Axis Name = No (the title of the chart should explain the two axes)
-
-
Y
-
Gridlines = No
-
Show Axis Name = No
-
Show Axis Values = No
-
Line
-
Plot
-
Plot Null Values = Yes (depends on data)
-
-
Series
-
Point Style = Outline
-
-
Values
-
Value Labels = depends on data
-
Font Size = depends on data
-
Pie
-
Plot
-
Value Labels = Labels
-
Label Type = Label - Value
-
Inner Radius = 80
-
Start Angle = 210 (but try other values to maximize size of chart within tile)
-
Single Value
-
Formatting
-
Conditional Formatting = Enabled
-
Delete the one rule that is automatically added - this will result in the tile title showing in black rather than a light grey which can be hard to read
-
Word Cloud
-
If you have purchased the Text Analytics package, the features there can provide more useful information than a word cloud.
Comments
0 comments
Article is closed for comments.