Domino visualisation in a column chart

This visualisation is brilliant. I have never thought about using columns in a column chart as domino pieces. One to add to your repertoire of visual tricks and let's hoop that this doom's day scenario does not play out. 

The Greek Bluff In All Its Glory: Presenting The Grexit "Falling Dominoes" http://t.co/wLGBOp30Sn pic.twitter.com/JB9BSQ0Xfr

— zerohedge (@zerohedge) July 4, 2015

Good and bad use of data charts

I came across these charts created by The National Geography in a special article about Food by the Numbers. The print article is an extraction from a video. (I made a poor quality image with my phone).

Like in most infographics, the rules of creating data charts are broken. It works well in the 2 graphics on the left, not so good in the map.

• The line charts abstracts away everything you do not need. Years are omitted, the scale is really rough. It focuses on the things that need communicating. The exact point estimate for "today" (2011), but still the number is nicely rounded (not 1,800,232,433). The growth trajectory is clear (without cheating with broken axes). Real data and projection are clearly separated.
• The connection to the pie chart is good. The pie chart itself is super minimalist with a huge data label to communicate and visualise the one number we need to know (could have bee "4%"). [Contradicting myself here, usually I do not like pie charts].
• The map works less well. The differences in size between the bubbles is hard to see. A simple stacked column might have been better here. Also, the really interesting statistic is to show the 4% broken down by region, not the absolute size. Maybe South America uses relatively more bio pesticides than North America?

---

Art: Pieter Breughel the Elder, the Harvesters, 1565

Graph screen shots

If you do not have access to the source data of a graph you have two approaches to getting the data into your presentation:

1. Use a screen shot

How to get the best looking chart from a screen shot:

• Make the chart as big as possible on your screen. Sometimes you can click the graphic which opens in a new tab in your browser. In PDF, you can zoom in without losing quality. Take the screen shot of this big graph and paste it big in your slide.
• Crop out all items of the graph that you can easily recreate in PowerPoint or Keynote: axis labels, chart titles, even the values on the X and Y axes. Next, recreate these items by hand in PowerPoint
• Cover as many elements on the chart as possible with a white box. Legends for example hardly ever look good. Cover it and create your own.
• Select the chart and pick "format picture" to see whether it looks better in black and white. Alternatively, use the colour picker to get your legend use the exact colour used in the chart.

2. Measure and recreate

When you do not use data labels in a chart (bar, column, line) but rely on a value axis instead, you can get a way with a lower of level of accuracy. You can literally print a chart out (the larger the better), measure the position of the data points and recreate the chart from scratch in PowerPoint or Keynote.

---

Art: Interior of a tailor workshop, artist unknown, 1780.

What to put in a short-term budget

In investor pitches there are usually 2 types of financial forecasts:

• Short-term: super precise, usually mostly cost
• Long-term: ball park, usually mostly revenue

Why do investors want to see some sort of short-term budget?

• While the long-term revenue outlook of a startup is highly uncertain, the near term cost drain is pretty much set in stone. Investors want to know where her money goes: in development, in an expensive, all-or-nothing/bet-the-company Super Bowl ad?
• Investors want to check the consistency of your story. You say that you are not focused on getting customers right now, so you should not be spending anything on marketing. On the other hand, you plan 25 new features in your product, where are the developers (and their salaries) who are going to make it happen

A short term cost budget does not need to contain 25 lines of Excel, by month. A simple x% of revenues number is a bit too simple though.

---

Art: Marinus van Reymerswaele, The moneychanger and his wife (1539), Museo del Prado, Madrid

The lone column

Most of the time, numbers in graphs look better than numbers in a table. There are exceptions though: when there is just one number, and when there is very little variation among the numbers. During my time at McKinsey, I have seen many examples of "lone columns", column charts with just one number in them, or tables full of tiny column charts with hardly any variations among them.

These charts are not only difficult to read, but they are also very hard to create in PowerPoint or Keynote SlideMagic's grid structure does it in a snap though, but hopefully users won't abuse the app for these type of consulting charts. Sign up for SlideMagic here.

---

Art: Painting of Trafalgar Square (c. 1865) by Henry Pether. Sign up for SlideMagic, subscribe to this blog, follow on Twitter

The only data charts you need

I have been designing presentation slides for over 20 years now (scary) and over time stopped using more and more types of data charts.

• Pie charts: I don't like the way they look, it has to place data labels, it is hard to compare two of them side by side
• Line charts have ver little presence (duh, a thin line) and I use column charts where possible to visualise time series. Yes, for correlations and hard core scientific data I might have to resort to them
• Clustered column charts, I find them confusing, it works better to just put 2 column charts next to each other
• Hybrid charts with 2 axes, very confusing. Again, I split them up into 2.

So, as a presentation designer you can get away with a very limited arsenal of data charts. Here is a quick run down of the ones I use: columns, stacked columns, and bar charts. (You can can guess which ones ship with SlideMagic)

The key to designing good data charts is careful, manual design (the opposite of copy pasting from a spreadsheet). What is the one single message that you want to pop out. What are the 10 to 20 data points that support this. Where to drop the accent colour, to what number of decimals should you round up the numbers. What breakdown categories should you group consolidate. Do I need a graph, or is it clearer to put the numbers in a single table? Data charts take time to prepare, but once you figured out what you want to show, can be produced in 5 minutes.

---

Art: Willem Claeszoon Heda, Breakfast Table with Blackberry Pie (1631), sign up for SlideMagic, subscribe to this blog, follow on Twitter

Design is detail

In management, being detail-oriented is not the behaviour that is considered good. Detail-oriented people get lost in tangents, loose track of the big picture, cannot focus to make decisions. Saying that you are not afraid of detail in a job interview will cost you points.

I think it all depends. Yes, staying stuck in unimportant tangents is not helpful, but when it comes to design, it is all about the detail.

You see this now best in mobile application user interface design. The screen is so small that you need to worry about every button or item you put in front of the user. I personally went through this experience when designing SlideMagic.

But slide design is the same. It is actually helpful to think of your slide as a visual on the screen of a mobile phone. This is sort of the perspective of an audience member who sits in the back row. Everything you put on the slide, everything, should be thought through:

• What words to use in the text box, can you cut more without losing the meaning, do you need to add more because it is too vague?
• The rounding of the data
• The order of the bars in the bar chart
• The order of the columns and rows in a table
• Are there duplicate messages? Does a text box say the same thing as the title?
• Do we need icons, or shall we call customers, well, customers?

All the detail will add up to a great slide that gives the big picture.

In case you wonder about the close up of Vermeer's painting "The Music Lesson", find out more here about Tim Jenison's attempt to recreate the Vermeer master piece using a lens, "accusing" Vermeer of being a very early photographer, rather than a painter.

Click here to subscribe to the blog