Data Visualization

I wanted to share a bit of work done and knowledge acquired at Coursera, offered by University of Illinois at Urbana-Champaign, offered at this link, taught by John C Hart. I really recommend anyone wishing to learn more and dig into the data visualization vast knowledge. This is a sample of the done on the second week, where the proposed challenge would be to create a graph on the temperatures averages found on NASA's temperature database. I have chosen to use Tableau as the tool for creating the graph.

The first decision is always how the most important information would be displayed, so I used a scatter plot to make sure the position of the dots describes the most important characteristic, that is to tell the history of whether or not the average temperature is rising since 1880. So the most natural way to display the information would be to have the cold / hot temperature in the y axis, and naturally the x axis representing the dates.

For reinforcement, the choice was to display the same information on more then one way, and an interesting way to achieve the goal is to add a blue/red scale to paint the dots, making the information easier to read. The color choices - lower temperatures in blue and high temperatures in red - are regularly used to express high and low temperature, so the graph is naturally read. I have also decided to plot all data for the reader to be able to look at variances within year, and use density to visually detect outliers.

A second goal would be making sure the minimum of ink is spent on anything that is not absolutely necessary to convey an important idea. So the horizontal lines are faded to an almost imperceptible level, while still hinting what the graph alignment really is. A peer has suggested that a color scale would help comprehension, but I did not include as the same information is portrayed by the x axis and I though it would pollute the result.

So, what are your thoughts on the work? How would you improve it?

Keywords: Data Visualization Software, Tableau Software, Data Virtualization, DataViz, Data Analytics