It’s time to talk about tables
Five situations where it makes sense to use an interactive table instead of a conventional visualization – and how to make them in Flourish
It’s time to talk about tables. Why, you ask? First, because today at Flourish we officially launched our new Table template, which makes it easy to create and publish interactive tables. And second, because we don’t do it nearly enough. (No, seriously!) For people who work with data, tables are ubiquitous – we use them every single day. What is an Excel spreadsheet if not a big, customizable table?
But more often than not, we only consider the table as a sort of intermediary form – a transitional state between the raw data and the finished visualization. We rarely think about it as a visualization form in and of itself. But it is, and I for one think it’s time we finally give the table some of the respect – and attention – it deserves.
It seems that part of the reason we largely ignore the table as a visualization type is that most of us are not sure when we should be using it. In an effort to clear this up, I did a bit of research, and I’ve come up with five broad principles to guide you in your table-making endeavors.
And instead of just telling you about these principles, I thought I’d show you an example that illustrates the instances in which you should in fact be using a table. Because I was using Flourish’s new template, making the table above was as simple as uploading a spreadsheet and tweaking a few settings. Click on a column to sort, or find a university by typing its name into the search bar.
Five situations to use a table
1. When you want to show individual, precise values
Rankings of any sort are a great example of this. In this specific case, you want to be able to look up where a university falls on the list. Sure, you might care if a university is in the top 10 or top 100, but what you really want to know is exactly what position it’s in.
The table above actually displays three different rankings for each university. As the default, the table above is sorted by THE rank, but you can sort Flourish tables based on any column. So if you want to sort by the QS or CUWR ranks instead, all you have to do is click on that column’s header cell.
2. When you want to compare pairs or small groups to each other, but not the entire dataset
For example, the first thing I did after compiling this dataset was check where my alma mater ranked in comparison to my mom’s and my brother’s. Did I care about comparing my alma mater to Harvard or the University of Toronto? No. Why? Because no members of my immediate family attended either of those institutions. Although it’s a relatively large dataset, all I care about is a very small subset of it, and the table format made it very easy for me to do that.
If this was a static table I would have had to look up each of the universities manually, but Flourish tables are filterable, so I was able to just search the name of all three universities and quickly compare them.
3. When you want to show both detailed and summary data
Alongside the rankings themselves, companies like Times Higher Education, QS and the Center for World University Rankings often release some of the information they considered when forming their lists. For someone reading this table to get information about a university, the fact that both the summary-level data (i.e. the rank) and the more detailed data (i.e. the number of students, student-to-staff ratio, share of international students and gender split) are displayed is really useful. And while most chart types are only capable of showing one type of data or another, tables – as this example illustrates – can show both.
With Flourish’s Table template you can even display collections of numerical columns as mini bar or line charts to make the more detailed data easier to read. In the university rankings example, a bar chart is used to visualize each institution’s gender breakdown.
4. When there are multiple units of measure
Most of the basic chart types are also only able to display one, maybe two, units of measurement – our university rankings table has more than five. It might make sense to do a scatter plot showing the relationship between two of the columns, like number of students and international student share. The only way to adequately visualize all nine columns is to use a table.
5. When you have a lot of categorical data
Imagine you had this university dataset and wanted to visualize just the names of the universities and all three of the rankings. How are you going to do it without a table? (The answer: Quite badly.)
You might try using a grouped bar chart, but even if you just wanted to show the top 20 THE-ranked universities, the scale is counterintuitive, the labels are too small, and there’s too much to read. It’s just too messy and too confusing.
We’ve done it – we’ve talked about tables
That wasn’t too bad, was it? I’m sure I’ve irked some data visualization experts with my incomplete and personal list of the reasons to use a table, so if you have a different set of reasons, let us know on Twitter. Or to get started making your own table using Flourish, check out the template page.