We Moved!

Many topics from my various blogs have been merging. So as the intersections of my mind grows, I merged my blogs together.

You can find all my blogs on www.onthemind.info now. 

I will be removing this blog in the near future as all content has been moved. 

Fundamentals of Pragmatic Data Visualization

Data Visualization (or DataViz as some people call it) is another of the catch phrases people use when talking about business intelligence. However, do you know what it is all about? What does it mean to visualize data? Why do you want to do it in the first place?

Let's take it down to the basics. In today's technology rich world, we are surrounded by data. To make sense of what is around us, we deal with data every minute. Whether it is the traffic lights while driving or a scatter plot showing the relation between profit and revenue; it is all about data. I talked about that in my earlier post "Data Data Everywhere... But Not a Byte of Information". 

So what is data visualization? Well, I define it as "The practice of showing data in a manner which highlights the message and leads to actionable decisions."

Let's dig deeper and identify the key building blocks or data visualization.

Data Data Everywhere... But Not a Byte of Information

Someone once said to me that we are generating data from the time we are born. In the life of a person, so much data is generated. Whether it is demographic data or transactional data or analytical data, there is so much of it that we generate every day. Really! You are generating data right now while reading this article. The site has registered you as a visitor... anonymous or recognized, that is upto you really, but the visit is registered into a datastore none-the-less. Depending on which site you visit, there may be information about advertisements and click actions that you perform. Now, I don't want to say that we are all being tracked and this is a conspiracy... it is a fact of life in these times. 

But what happens with this data?

Is your data beautiful?

Some people say, "Beauty is skin deep". Others disagree with that statement and postulate that there are many factors that make someone beautiful and it is not only the visual appearance.

This is true about data visualization also. There are people who think of visualizations based on what they see visually. When it comes to the data though it is quite clear, in my opinion, that good visualization (read: beauty) is not just about what you see, but what the viewer understands from it.

The visualization you choose is as important as the data being shown. Let's take two very public and common examples showing the same type of information. These visualizations show country based information about different activities in the internet.

1. The "Bubbles":

This visualization shows the countries that viewers of your LinkedIn profile belong to. The visualization shows each circle representing a country. The size and color density of the circles here show the relative number of visits from that country. For the smaller circles, the user has to hover over the circle to understand the information about the country and the number of viewers. 

This may look like a neat way to show this information, with the country where your largest visitor base is being at the center of your universe (i.e. your focus point). However, is it a good visualization? Does it pass on the information that it is representing? 

In my thought, the answers to those questions is "no". The visualization as problems at multiple levels. First, it is not easy to understand that the circles mean countries. If I want to highlight my largest visitor base (i.e. center circle), shouldn't I also mention the country that circle represents? I have a name for the second largest country. The smaller circles really don't mean anything. A couple as you see are just dots.

Overall, maybe a "neat" way to show the data, but not "beautiful". 

2. The "Map":

Google Analytics on the other hand shows country data in a map visualization. This visualization shows a map of the world. The density of the color indicate the relative rank of the countries based on visitors. The darker the country shading, the more visits from that country. 

This chart does show me all countries and takes my eyes to the countries which are darker shaded to focus on those. However, like before, is it a good visualization? Does it share the information that the user is looking for?

Perhaps "not". The visualization is surely better than the "bubbles", since now I can see all the countries and quickly understand the ones that are important to me (i.e. the darker shades which stand out). It is also intuitive as I can see that the data segmentation is based on geographic locations (i.e. countries). But that is about all it does. It does not tell me how much better one country is over another. Given the size of the map, it is not able to do justice to the smaller countries, ones that are not big enough to show up in the map.

Overall, more beautiful than the bubbles, but still does not show the full beauty of the data.

From these two public visualizations, it is clear that they do not do justice to the underlying data. Lets look at one that does.

3. The Simple Bar Chart:

Consider a simple bar chart with countries listed as categories, using the length of the bars indicating the amount of sales. This visualization example is easy to understand by the user. It is clear that some countries have more sales than others. It is easy to sort it and show ranking with the top countries at the top or bottom, depending on what you want to highlight. You can also add some more data values to show the numbers on the chart, though that may not be always needed.

Is this a visualization that does justice to the data? I think so. It shows the data for what it is. It shares the complete beauty of the data and brings that across in the visualization for the viewer to quickly grasp and make decisions on.

Incidentally, both Google Analytics and LinkedIn use similar bar charts to show other data about visitors. Which look spectacular. I wonder why they chose such mediocre visualizations when it came to country based data.

What are your thoughts? I would love to hear from you about these and other visualization examples. Do you think "Beauty is skin deep", or does the statement "There is more than meets the eye" more accurate according to you?