“The greatest value of a picture is when it forces us to notice what we never expected to see.” – John Tukey.
Whenever we analyse data, our goal is to highlight it’s features in order of their importance, reveal patterns and simultaneously show features that exist across multiple dimensions.However when we visualise data, there are a few things that we have to think about.
We have to think about the amount of data we are planning to use and the level of information we are hoping to portray. We have to think about whether this kind of information changes over time and thus collate data the can be used to witness these changes over time.
But most importantly, we need to answer a question. There has to be a meaning behind what we are asking and what it means for us to put this visualisation out into the world.
“Why was the data collected, what’s interesting about it, and what stories can it tell?” Why is this data meaningful and who is it meaningful for?
By focusing on the original intent of the question, we can present something that provides a benchmark for what is and what is necessary, as well as provide insight to other questions that need to be answered.
“A visualisation should convey the unique properties of the data set it represents”
Each data set is different and so the point of creating a visualisation is to “expose that fascinating aspect of the data and make it self-evident”.
This means that we need to use the right type of visualisation, the right colours and provide the context for which the visualisation is important. It must be easily understood as it stands as well as providing a sense of mystery and wonder.
Should the data set be visualised?
If the data set benefits from being visualised because the dataset contains complex elements that cannott be shown in an excel spreadsheet or as a bar graph, then it should be visualised in a way that someone can undertand it’s concept and how it relates to the world.
“Just because it can be measured, doesn’t mean it should”
This is important. Just because we can, doesn’t mean that we should. If the dataset provides no meaning and there is no mystery to what it could mean, then there is no point in visualising it. Again, a question needs to be formulated before deciding to create a data visualisation.
[Reference: Visualising data – Ben Fry]