Guidelines for a better data visualization

With the arrival of big data, and the reliance of decision making on data that comes at us with overwhelming velocity and size ,that we cannot comprehend without some form of abstraction, such as a visual one. Ware (2000) stated that visualization is one of the important tools for effective research presentation and communication due to its ability to convey and synthesis big data into effective graphs. A picture worth more than a thousand words, it’s easier for our brain to understand and comprehend pictures than words and numbers (Cukier, 2010).The increasing size of the data obtained and the easiness of accessing such data require the introduction of effective ways to interpret and analyze the information in a simple, easy to communicate format (Cukier, 2010; Szalay and Gray, 2006).

When going through the reading, I could not help but thinking about if there exist a set of principles or guidelines that will be easy to follow for a better and more effective data visualization. I came across an article by Christa Kelleher and Thorsten Wagener (2011 ) that have ten guidelines obtained from books and journal articles that I thought will be helpful for our class to read

1. Create a simple graph that will explain and convey the message that you want your audience to understand (Tufte, 1983 [pp. 91–137]), avoid redundancy in your graph as that might overcomplicate the plots purpose (Tufte, 1983 [p. 93]). while insuring the reader can still distinguish between the different aspects of visualization such as shape, color, thickness (Cleveland, 1984).

2. Think about what type of encoding objects and attributes; such as points, lines, or bars; will be used to communicate certain pieces of information in your plot, because humans can interpret and quantify certain plot attributes such as 2D data (length, position) better than others (Cleveland and McGill, 1984).

3. Based on the reason or purpose of your plot focus on visualizing pattern or visualizing the details of your data (Few, 2004a; Kosslynand Chabris, 1992). For example if your plot will be representing a pattern a heat map or a bubble plot will be effective. A bar or line graphs when individual data points are important.

4. Select an axis range that will not mislead your reader or misinterpret your data (Robbins, 2005[pp. 239–241; 285]; Tufte, 2006 [p. 60]; Strange, 2007 [p. 89]). For example, selecting a range for the Y axis depends on the graph’s purpose and type, if you are representing absolute magnitude then the axis should start at zero (Robbins, 2005 [pp. 239–241]; Strange, 2007 [p. 89]).

5. Be considerate when transforming your data from one scale to another (Cleveland, 1994 [p. 66, 95, 103]), an example, is data transformation from a linear plot to a log plot. Data transformation should be dependent on the data set and the purpose of the graph. The wrong transformation might cause the misinterpretation of data.

6. If points are overlapping make sure density differences is easily seen in scatter plots (Few,2009 [p. 121];Cleveland, 1994 [p. 159]). For example changing plot points from opaque to transparent enhances the message or information presented by easily seeing the density difference.

7. When plotting time series information, use lines to connect sequential data (Strange, 2007 [p. 150]). For example if a plot connect non sequential data on either side of a period of no data available or missing data, this might cause the reader to interpret this line as a linear change between those data points.

8. Simplify and plot large data sets by using meaningful ways (Cleveland and Devlin, 1980; Chambers et al., 1983 [pp. 21–24]; Cleveland, 1994 [p. 187]). For example use box plots to simplify large data sets.

9. When comparing variables or values, keep axis ranges as similar as possible (Cleveland, 1994, [pp.86–87]; Few, 2009 [p. 180]). When using the same axis ranges, data will be better interpreted.

10. Select the color scheme based on the data plotted (Brewer, 1994; Harrower and Brewer, 2003). If the color scheme chosen matches the type of data used, this will further support the visualization and interpretation of the data plotted.

The above guidelines are presented in fig. 1(a,b) and fig. 2

References

Cukier, K., 2010. A special report on managing information. The Economist 394

Szalay, A., Gray, J., 2006. 2020 Computing: science in an exponential world. Nature

440, 413–414.

Tufte, E.R., 1983. The Visual Display of Quantitative Information. Graphics Press,

Cheshire, CT.

Cleveland, W.S., 1984. Graphs in scientific publications. Am. Stat. 38 (4), 261–269.

Cleveland, W.S., McGill, R., 1984. Graphical perception: theory, experimentation,

and application to the development of graphical methods. J. Am. Stat. Assoc. 79

(387), 531–554.

Few, S., 2004a. Eenie, Meenie, Minie, Moe: Selecting the Right Graph for Your

Message. Intelligent Expertise. Perceptual Edge. Available at <http://www.

perceptualedge.com/articles/ie/the_right_graph.pdf> (accessed on 3.08.10.).

Kosslyn, S.M., Chabris, C.F., 1992. Minding information graphics. Folio, 69–71

Robbins, N., 2005. Creating More Effective Graphs. Wiley-Interscience, Hoboken, NJ.

Tufte, E.R., 2006. Beautiful Evidence. Graphics Press, Cheshire, CT.

Strange, N., 2007. Smoke & Mirrors: How to Bend Facts & Figures to Your Advantage.

A & C Black Publishers, London, UK.

Cleveland, W., 1994. The Elements of Graphing Data, second ed. Hobart Press,

Summit, NJ.

Few, S., 2009. Now You See It. Analytics Press, Oakland, USA.

Cleveland,W.S., Devlin, S.J., 1980. Calendar effects in monthly time series: detection

by spectrum analysis and graphical methods. J. Am. Stat. Assoc. 75 (371),

487–496.

Chambers, J.M., Cleveland,W.S., Kleiner, B., Tukey, P.A., 1983. Graphical Methods for

Data Analysis. Duxbury Press, Boston, MA.

Brewer, C.A., 1994. Color use guidelines for mapping and visualization. In:

MacEachren, A.M., Taylor, D.R.F. (Eds.), Visualization in Modern Cartography.

Elsevier Science, Tarrytown, NY, pp. 123–127.

Harrower, M., Brewer, C., 2003. ColorBrewer.org: an online tool for selecting colour

schemes for maps. Cartog J. 40 (1), 27–37.

Ware, C., 2000. Information Visualization: Perception for Design. Morgan Kaufmann

Publishers, San Francisco CA. (8671), 3–18.

C. Kelleher, T. Wagener.,2011. Ten guidelines for effective data visualization in scientific publications ,Environmental Modelling & Software 26 (2011) 822–827

ins”: fQ���P�