top of page
  • Writer's pictureT. @ Data Rocks

No, it doesn’t need to be a bar chart, featuring Show Me the Numbers by Stephen Few.


When I was starting out with data, Excel was my safe space. I learned to love it. Needless to say, for a good chunk of my career, my understanding of data was limited to what Excel allows you to do. While it can be an extremely versatile tool, if you didn’t have any previous knowledge of working with data or visualisations, you’ll too believe that the standard offers in the tool are ok for what you need to achieve. But, as I advanced and got more experience and feedback on my work, I started facing some challenges.


Notoriously, tools like Excel offer some questionable standard options when it comes to depicting data - all labels enabled, lots of unnecessary colours, and titles everywhere. A lot of thought needs to go into what makes a chart communicate data effectively. The devil is always in the details, and it can be tricky to navigate all of these small choices without guidance.


If you’re starting out in data and you’re taking an interest in Data Visualisation, chances are you came about several book recommendations. You’ll find countless listicles, and in all of them, you’ll surely find at least one of Stephen Few’s books.


His first book, Show Me the Numbers: Designing Tables and Graphs to Enlighten, was released in 2004 and became a classic in the field. I own the second edition, published in 2012, which contains a bit more material and goes further in depth in some subjects. This is the version my review is based on.



Tables are Data Visualisation Too


If you’re like me, you spend most of your days creating visuals for corporations. In corporate business, accountants and finance types abound, and there are some obvious preferences when it comes to representing data to a certain profile of busy executive. Cue tables and some variation of a bar chart. Sometimes a trendline. But tables. They love tables. They crave tables. Serve them a graph, and the question will inevitably arise: “can I have this on a table?”


Familiarity certainly plays a role, and for a long time, this was all everyone had: long and colourless tables. Then came more and more data. And finding meaning in ever-growing tables became a cumbersome task. That’s when data visualisation started getting traction with corporations. But transition isn’t always easy, even when it’s necessary - so we’re still stuck with using tables and common chart types such as bars and lines in most of these environments.


The great contribution of Show Me the Numbers by Stephen Few is to be a book written for those of us who have to navigate the communication of complex data within a business environment. It serves as a good starting point and guideline if you’re just starting out and a bit unsure what all of this datavis talk is about.


First, Few introduces a bit of the basic statistics we all have to be aware of when dealing with KPIs, goals and metrics - medians, averages, aggregates, and standard deviations. He doesn’t go in-depth, but it is a good primer if you’re new to the field and need a place to start. Few also focuses a lot of his guidelines under the assumption that datavis means the representation of quantitative data. Then, he takes us on a journey about the most fundamental and common ways we represent data in a business context: tables, bar charts and line charts. This is one of the rare books about datavis out there that extensively focuses on designing better tables.


And how the author likes them! He offers a unique framework of how tables work - as simple as they may seem - in the depth of detail. Show Me the Numbers is indeed about showing numbers - arranged primarily in rows and columns. This is one of the hallmarks of the book: it appreciates detail.


The book is also a good set of guidelines and rules about data visualisation, which is appreciated primarily by those just starting out. If you’re looking for some brief, straight-to-the-point guidance on improving the design of your tables and bar charts, this book has got you covered extensively.


Here are some of the lessons I took from Show Me The Numbers that I use in my work:

  • Clarity is essential: the primary goal of data visualisation is to communicate information as clearly and effectively as possible.

  • Understand the audience: different audiences have different needs and levels of understanding. A good datavis should be tailored to meet its specific audience’s needs.

  • Choose the right chart: different chart types are suitable for different kinds of data and different questions. Line charts are best for time series data, bar charts for comparisons, scatter plots for relationships, etc. I should try to depict data in the best possible way, considering the message and intentions of the visual.

  • Simplicity is key: extra decorations should be avoided. A clean, simple design is often the most effective way to communicate data.

  • Design with purpose: every element in a visualisation should be there for a reason I can justify. If it doesn’t, it’s likely contributing to noise and distraction.

  • Highlight the important features: carefully and purposefully use visual cues (like colour, size, and position) to highlight the most important parts of the data.

  • Maintain integrity: always strive to preserve the integrity of the data. Avoid misleading representations or distortions. Data visualisation aims not just to present data but to do so in a way that is truthful and fosters understanding.

  • Numbers need context: Numbers without an appropriate level of detail and context around them have no meaning and won’t fulfil their purpose of informing others about the data being represented.



“Silly Graphs That Are Best Foresaken”


This is the title of chapter 12 of Show Me The Numbers. As much as I appreciate the rest of the book’s content, and as much as I often go back to it for reference in some aspects of datavis, there is a running thread throughout the book that never quite sat right with me. Stephen Few has very strong opinions about what is considered good data visualisation. His view is mostly formed based on business and academic reporting. When I started creating data visualisations, and that was all of my exposure to the world of data communication and information design, I tended to agree. But as soon as I stepped outside that context and learned of the incredible ways we can creatively communicate a data story and inform others through visual devices, my view of the “rules” of datavis changed dramatically.


Show Me The Numbers is prescriptive. It is a guide. It tells you how things should or should not be done. The issue here is that technologies evolve, and practices shift. What is deemed unacceptable today may be the biggest breakthrough tomorrow. It also comes from a place where simplicity and speed to insight are paramount. It presents datavis as a tool to display information to busy business people. It doesn’t go into other purposes datavis can have. It doesn’t consider the use of visualisations as means to persuade, engage, entertain or delight. This approach is to datavis like brutalism is to architecture - it is not bad or wrong, just a bit too bleak and utilitarian. And while it may be true that most things could be more quickly and simply explained with a bar chart, those are not always the only objectives of a visual display of data.


One of the probably unintended consequences of being seen as an authority in a field and writing about how every round-shaped chart should be eliminated from existence is that others will repeat these postulates. Imagine if you were Florence Nightingale, concerned about the lack of sanitary measures during a war, and the main comment on your carefully crafted visuals was, “this should’ve been a bar chart or a table”. If she was to choose one of the more conventional forms to display the information she had, would it have been as effective? Would she have caught the attention of those who could influence change? Would we still talk about them today? I’m glad nobody forsook the Nightingale Rose just because it relies on a round axis.





Atypical or novel types of charts may be a bit more challenging to read at first contact. But there’s a lot to gain as a field if we embrace them. Some of them will work, some of them will not, but we’ll never know if we don’t try. Bar charts were new in the 18th century too. There are wonderful pieces of work from designers like Shirley Wu, Nadieh Bremer, Stefanie Posavec and Giorgia Lupi that challenge our perception of “appropriate” chart design and are not only effective but also engaging and beautiful.


Will creative data visualisation make its way into business reporting? My experience says yes - my supply chain clients love a Sankey chart - but it may be a journey we have to start, as designers, exposing our clients and stakeholders little by little. It also comes from a desire to engage as well as inform. Until very recently, engagement and adoption weren’t as big of a deal when dealing with information - but as technology changed, so did our attention spans. Sometimes, all you need is an interesting chart type to lure people in.


By the way, have you ever thought that analog clocks are pie charts in disguise and that many of us read them intuitively with no problem? Food for thought.



Should you read it?


If you’re looking for a great guide to get started with basic chart forms and learn a lot about good design techniques to create detailed views of data, this is a good pick.


It is a classic of the field, but it is basic and prescriptive. Read it carefully and with a critical eye - not everything needs to become a bar chart, and you don’t need to hate pie charts just because someone said so, to a corporate crowd, 11 years ago.



Always check your local library first to see if any of the books I recommend are available. If they’re not, consider donating a copy!

 

Get a copy at your local library | Amazon


 

If you subscribe to my monthly Newsletter, you’ll get a summary of all recommendations, plus more of my data viz musings.


You can also follow Data Rocks on LinkedIn or read this and other articles on Medium.

 


196 views0 comments

Comments


bottom of page