Last Wednesday I had the luck to contribute with a presentation to the “IV Jornadas de Gestión de la Información” (4th Workshop on Information Management”, organised by SEDIC the Spanish Society for Scientific Documentation and Information. My talk was a little tour around Information Visualisation explaining the most emblematic applications along with some of the most advanced ones, capable of changing the rules of the game. 

These aren’t, however, the most appropriate for immediate application in the day to day practice. Hence I had the sensation that, despite the apparent success of my speech, I didn’t stress enough the fact that the creation of the most modest and usual graphics that we see in some presentations is not only an important part of this discipline, but the one we can all directly benefit from.

For this reason it’s interesting to reflect how to increase the effectiveness of the graphics and diagrams we use in our presentations. Spreadsheets are a common tool in today’s work. Every user of one of these tools can also build graphics based on the data, but the use of them is still precarious in many organisations.

In issue number 73 we discussed the general process of building a graphic, but before doing this we have to first answer the following question: when should I use a graphic? Edward Tufte professor at Yale considers 5 ways to show quantitative information. We will reduce them to 3 defining them as follows:

• Sentence: In this context it is a sequence of words expressing the relation between several data. Conventional sentences aren’t useful to show more than 2 or 3 data, since they don’t allow you to compare them in a simple way. See the example in issue number 45 "Informing Visually"
  
• Tables: A table is an array of data ordered into rows and columns, clearly separated from the surrounding text. Typically a table is an ideal resource to show exact number if it contains between 3 and 20 data. With a higher number of data it’s advisable to consider the possibility to use graphics. This doesn’t mean that some bigger tables aren’t an excellent way to show the data (the balance sheet typically has more than 20 data and I haven’t yet found a chart that makes it more understandable, specially in low profit periods…)
  
  
• Graphics: An artefact of visual nature generally represented on a flat surface, devoted to expressing relations between data. It's especially useful for large amounts of data.

Tufte also distinguishes between “table text” and “semi-graphics” that we consider as a combination of table or graphic with sentence. It is made of one of these resources surrounded by text; see “The Visual Display of Quantitative Information”, chapter 9

So, the core of today's issue could be summarised into a sentence:

It's advisable to use sentences for relations between 2 or 3 data at most. The use of tables is recommended between 3 and 20 data and graphics from 2 on. You will get the most of graphics the larger the amount of data. 

But, in order to illustrate the message we want to transmit, we could use a table as well.

But we could still use a graphic to express the same information (see the figure)

¿When should I use a Graphic? This representation, in logarithmic scale gives an idea of the scope of sentences, tables and graphics as a function of tha amount of data and / or relations to represent. The darkest the colour the most appropriate is the use of the corresponding category. The three options overlap. The gradation of colour indicates that theres no unique decision nor well defined limits. Click on the image to enlarge it.

Colour gradation indicates that neither 3 nor 20 are definitive numbers, it depends on the application. The graphic adds in this case an extra expressivity to the exact number of the table and reduces the verbosity needed to explain this message. You should judge by yourselves which one among the three is the most appropriate for this case.

Then, why this long article if the contents can be summarised in a brief table or an elementary diagram?. Because context, a key element in order to convert information into knowledge, and nuance are difficult to transmit by means of tables and graphics, even when we are talking about quantitative information. 

The same happens with personal experiences and opinion, undividable parts of the articles you have been kind enough to read. For the appropriate combination of the different elements is essential to gain a better insight.

Links of this issue: