Ultimate Guide to Tables in Tableau
Part 1: The Introduction
Quick without thinking too much about it answer the following questions:
- What chart type is used the most when displaying data?
- What chart type is the hardest to design?
- What chart type is the least customizable?
One thing is for certain: it’s not just the savvy analyst who thinks tables are soul-sucking to create: almost all digital-centered designers think the same too. But here’s the reality: people use tables. They use them because it informs their decision making. It’s our job as practitioners to provide them with the tools that allow them to do their job. If that means making tables, then we should make tables!
We’ve started this series because there is no definitive source for helping to create tables in Tableau. Rather than wait for product improvements we’ve decided to take tables head-on. Over the course of this series we will show you different ways to improve the design of your tables so that they are more engaging for your users. By the end of this series you will be building dynamic tables that highlight key insights while retaining the format of a table. This first post goes in-depth on how experts in user experience and user interface think about creating tables. From there we’ll see exactly what’s possible in Tableau, identify gaps, and hopefully create new solutions.
A quick search of the web on designing data tables will point out one thing: designers have spent a lot of time thinking about how to design tables. In fact, thousands of articles to sift through.
So what makes a great table?
Principle #1: The purpose for the table is extremely clear
It’s easy to provide a table of data for any audience, it’s another thing to figure out the exact needs of the audience and then tailor the output of the table so that key insights are highlighted while additional data provides context. Maybe that key value is total customers. Maybe it’s a change in customer from the previous month. Maybe it’s the change in total customers from the previous month compared to the same change a year ago (delta of a delta). How we develop tables with customer data will look different for each of these three objectives!
Principle #2: Tables need to be extremely flexible and adaptive
How your audience utilizes tables will differ from person-to-person. To allow users to gather their own insights, it should be easy to:
- sort rows
- re-order or add columns
- resize columns.
- freeze row and column headers.
- change the density of the number of rows displayed.
Principle #3: Tables need pagination
Scrolling is great but consensus among designers is that tables should have a fixed height. This allows the designer to control the entire user experience.
We’ll come back to this one as a separate article in our series.
Principle #4: Tables need filters (but filters shouldn’t be the star)
If there is one thing we’ve all encountered when developing anything in Tableau is that our audiences want to filter for EVERYTHING. This ultimately leads to a giant column or a long row of filters–I’ve personally seen a dashboard with 50+ filters on it before, and, yes, it was extremely ugly and took away from any insights and performed terribly on Tableau Server.
We’ll also come back to this one as a separate article in our series.
Principle #5: Tables should be able to perform bulk actions
We should be able to select many rows, or individual rows to perform bulk actions that update a table. While Tableau is primarily for displaying data, it is possible to edit the data with write-back functions using extensions or, for the hacky, using parameters.
This one tip was consistent amongst designers so we’ll try to come back and discuss this in detail.
Principle #6: Tables need context.
Tables should provide their audience additional information at all levels. At the global level its showing KPIs or other charts prior to diving into a table.
At the row level it’s the ability to “click into” an individual row to understand the meaning of the values in the table. This might mean having a complex tooltip, a slide-out container, or expanding rows that provide additional information.
This also means values in the table might have sub-text, icongraphy, imagery, or status indicators associated with individual values.
And for columns of data this means using selective color. Color should only be used in two-to-three columns of a 10 column table. For tables with less than five columns should only use a single color.
Principle #7: Tables formatting needs to maximize readability.
Despite the notion that tables are simplistic in their ability to display information, they are actually complex systems of information that must be formatted with great care to ensure their audience can maximize understanding.
The reminder of this post will be dedicated to the components that help maximize readability of tables.
Let’s start with a table we are all guilty of making at some point in our careers: the small-font, information-dense table.
In this example we’re not thinking at all about the end-user or how we could design the table to improve decision-making using the table. Our only thoughts are: how can we cram information into this table so I don’t have to think about it ever again. But tables can be useful. And we shouldn’t care about space because we basically have as much space as we want; it’s digital, it’s not paper, so don’t be afraid to use pixels for good.
So what are some easy things we can do to maximize readability of our tables?
Tip #1: Make font sizes readable.
Don’t make your font sizes small. A novice table creators inclination is to cram into a table by minimizing font size. I find that a font size between 10 and 14 is more than appropriate.
Tip #2: Distinguish headers from the body.
Use bold fonts or changes in color of the header text to distinguish the header from the body.
Tip #3: Add spacing and padding to rows in the table.
Row padding should be between 0.5 to 1.0 times the font size both above and below the text.
This means our table expands from a height of 300 pixels to 500 pixels. The expanded real estate means there is more room to focus on the numbers.
Tip #4: Choose table-friendly fonts.
Be thoughtful about your font selections. I personally avoid serif fonts and choose either sans serif or mono fonts. Overall it’s important to choose a font where numeric values are monotonically spaced. This will allow digits to align across rows in a column.
In the next posts we’ll dive into each of the different key principles.