Data visualisation has blossomed into a multidisciplinary research area, and a wide range of visualisation tools has been developed at an accelerated pace. Preliminary statistical data analysis benefits from data visualisation to form the basis for decision-making. There is a greater need for people to make good inferences from visualisations. The flexible nature of current computing tools can potentially have a major impact on the learning and practice of the discipline of statistics and allow easier use of visualisations in the educational process. While this view has many merits and we support its general spirit, we argue for a valuable role for a non-visual approach at certain points. Students will employ data visualisation in an OPEN Data context. This chapter is a theoretical discussion of a framework, which emphasises explicit assumptions that help to direct inferences appropriately. In particular it addresses the common illusions of causality in student reasoning. Our discussion of points of disagreement is based on specific theoretical concerns.
TopData Visualisation
Data visualisation allows us to explore and effectively communicate relevant information about voluminous data through graphic representations. This graphic mode includes visualisation of all kinds of information, and is closely associated with research by computer scientists. From the perspective of statistical pedagogy, data visualisation can be viewed as including computer-assisted exploratory data analysis of voluminous complex data sets. In this paper we do not use the technical distinction between BIG Data and OPEN Data. We only use the term OPEN Data given our expectations of what would be used in a classroom. The new large resources of OPEN Data, such as offered by several official statistics agencies, and any of their accompanying Online analytical processing (OLAP) facilities, will be relevant to the arguments and claims of this paper. The emphasis we offer is on the coherent learning experience that can be marshalled from these resources.
Developments in computing power have greatly enhanced graphical representations in recent years. Computing advances have enabled the drawing of precise, complex displays with great ease, making rich graphics drawn for the purpose of illustrating and explaining results and relationships widely available. Moreover, computing advances have also strengthened exploratory graphical representations that in turn provided support to exploring data further. The quality and quantity of graphical representations has been improved. Many distinct displays of the same data can be drawn to shed light on hidden aspects of the information. It is important, however, to point out that the real revolution is not in the machines that calculate or provide various displays of data, but in the ways we use the data.
The advantages of graphical representations of data have been gradually appreciated and capitalised on (see, e.g., Tufte, 2001; Gal, 2002; Espinel, Bruno, & Plasencia, 2008; Batanero, Arteaga, & Ruis, 2010). These advances provide an opportunity for new avenues of education for young students of statistics. The great importance of computer software availability and associated popularity, in determining what analyses are carried out and how they are presented, is a topic of major relevance. In the world of business, the spreadsheet Excel has long been in common use for creating graphical representations of data. In the world of statistics, there are several sophisticated software packages used by statisticians, including SAS, SPSS, STATA, GENSTAT, STATISTICA, S and S-PLUS, and more recently R.
A great mass of data is being continually generated. Citizens of present and future eras will be constantly and increasingly bombarded with tables of data and graphics in media, economic forecasting, and social activities. People’s decisions about their everyday lives depend increasingly on data visualisation and numerous reported figures, and these representations are not always what they seem: “There may be less in there than meets the eyes” (Huff, 1954, p. 8). Although Huff wrote this remark several decades ago about graphics used in the print media of his day, his ideas are still applicable in visual media.
It is essential to empower citizens with the skills to become informed about our world and understand the data that politicians, advertisers, and other advocates are using to promote particular causes that will impact the future of our planet. This prerequisite for citizenship embraces not just the capacity of understanding the underlying messages that the data potentially reveal; citizenship also includes critically examining statements presented in news media in terms of data and data representations, so as to recognise the varieties of misleading or distorted visualisations that miscomprehend and misrepresent data.
Misrepresentations of data may be created intentionally to obstruct the proper interpretation and to induce desired but incorrect conclusions. They may arise accidentally, for example when users are unfamiliar with using graphical software, when the data cannot be accurately conveyed, or when implicit or latent assumptions have been simply imposed on the data and inferences. Whether the deception is intentional or not, being able to recognise such distortions is a skill to be valued in citizens: