# Make better business decisions with Visual Analytics

People often say ‘a picture is worth a thousand words’ or ‘you have to see it to believe it’. It seems that everywhere we turn right now this is even more true. We see information pictured in infographics and clever animations and I have to ask myself why? And why now?

I have spent some time considering why visualisation is so hot now and why it may be important to the overall process of business decision making. My quest to understand led me to ancient history, where historical events and future prophecies were depicted in hieroglyphs.

The graphical representation meant everyone could understand, information was available to people of all levels.
Now jump forward to 1977 when John W. Tukey in his work ‘Exploratory Data Analyses’ said

“The greatest value of a picture is when it forces us to notice what we never expected to see”.

In business we can still learn from the ancients:

• We still need to understand the past so we can enhance our business decision making
• We still need to predict the future to avoid pitfalls and seize opportunity

Let’s illustrate with a real example. We will use some data developed by Francis Anscombe in 1973 to show the value in visualising before deciding, I hope you find it as demonstrative as I do.

Firstly, look at the data, a set of numbers, in all four groups of numbers (X1Y1, X2Y2, X3Y3, X4Y4) summary analysis of the data (traditional BI) would say:

 Statistic X Y Sum in each case 99 82.5 Mean in each case 9 7.5 Variance in each case 11 4.1

And the measure of correlation between x and y in each case = 0.816 indicating a very strong relationship in the data.

You can see in the table below how Anscombe's data looks in a tabular format,  you can only draw one conclusion, that all the set's of data behave the same way, so one business decision will apply to all the data I have -  Job Done!

Using SAS® Visual Analytics I did a simple autochart on these data points to illustrate the value of visualisation, see below:

I have now discovered that in fact I may have easily drawn the wrong conclusion using simple statistics or traditional small data BI – imagine that’s a key decision about risk, finance or the customer? So, Job Not Done! I need to do some further analysis, I need to understand more about what's happening before I make an uninformed business decision.

The Analytics Lifecycle

It’s clear that the visualisation aids in the process of discovery. Discovering what the data is telling me and in turn how to analyse my corporate data to gain the most insight and benefit.

Visualisation forms a key part of the decision making process and the overall analytics lifecycle. The Analytical lifecycle describes the overall way in which organisations implement an analytics strategy. Visualisation contributes in key areas including:

• Identification and formulation of the problem.
• Data preparation.
• Data exploration.
• Evaluation / Monitor Results phases.

Why should you seriously consider visualisation as a key part of your big data analytics strategy?

Visualisation provide benefits to everyone:

• Allows everyone to contribute to the decision making process no matter where they are.
• Allows the creation and sharing of insight FAST on practically any amount of data.

Take a look at it now - try it for yourself for free right here.

1. Posted February 12, 2013 at 3:05 pm | Permalink

Visualisation at the Data Exploration stage is imperative as you have shown with the Anscombe example.

What I love about this sample data is that the correlation statistic is the same 0.816, so someone may assume there is a strong linear relationship however from the scatter plots it is clear that there other types of relationships happening. Without visualisation you can easily miss seeing the true relationship and influential observations.

• Posted February 12, 2013 at 3:23 pm | Permalink

Thanks Michelle,
Agree, I think this also points to the importance of understanding the process of analysing data. EDA (Exploritory Data Analysis) or as people call it these days Data Visualisation is a key technqiue for avoiding what can be serious mistakes in decision making.

With Visualisation you can "Look before you leap".

Thanks for the comment, always sappreciate them and it show's people are reading :-)

• Posted February 12, 2013 at 3:51 pm | Permalink

And perfect timing with the blog post as I show the attendees on the SAS Business Analytics bootcamp this week :-)

2. John Kershaw
Posted February 24, 2013 at 8:39 pm | Permalink

But this type of analysis has been around for years, there's nothing new in plotting your data? Tukey's book was published in 1977 and any statistician worth his/her salt should have been doing this type of analysis regardless of something new called 'visual analytics'! Maybe this will reinforce the need?