Abstract:

In subjects of statistical contents, the students need to understand when and how they must use a number of techniques and procedures for the analysis of a data set. In this respect, the teaching activities must be designed so as to develop the students’ skills for an appropriate application of the referred methods and also an adequate interpretation of the results provided. Different computer programs can be used to accomplish the latter tasks, depending on the application setting and the available software in each University. The difficulty in finding appropriate statistical computer tools increases in the post-graduate studies, as they involve the application of more complicated approaches requiring specific software, which is typically more expensive than the standard one.

We will focus on the process of learning and teaching the theoretical contents of a particular subject, referred to as Spatial Statistics and included in the Master of Statistical Technics in the University of Vigo (Spain), although the ideas developed in the current report can be extended to other postgraduate statistics subjects. The theory of Spatial Statistics has been designed to deal with data where the spatial component is known and provides relevant information for the study, so that the spatial reference must be included in the analysis. Different application programs are available in the market to handle spatial data, whose cost usually depends on the number of licenses acquired, allowing the users to characterize the spatial dependence and also to obtain predictions at unsampled locations. Cheaper alternatives are based on self-creating the required software by applying any of the programming languages, where even free-cost alternatives are available. However, the latter options often require supplying support on their use to the students, so that these tools must be preferable for practical lectures rather than for theoretical ones.

In our experience, the preliminary introduction of concepts and techniques of spatial statistics asks for the use of simple application programs to illustrate their meaning and implementation in a friendly way. In this respect, we propose the use of Excel for the latter purposes, because of its simplicity together with the wide use and availability of this software, which is in general known by the students. With this aim, ad hoc templates can be designed to proceed with different approaches, where the necessary steps are described, and even sketched, but need to be completed to provide the users with the intermediate and final results. The required work lets the teacher exemplify the performance of a particular technique and, at the same time, the students develop the same implementation and put into practice the desired approach, which is even applicable for long-distance learning, when the students attend lectures through a video-conference system. In this contribution, we will present the templates prepared for the following issues: approximation of a nonparametric semivariogram, implementation of a parametric semivariogram from a nonparametric one, application of the ordinary kriging technique and implementation of the nonparametric prediction approach.

Keywords:

Computer program, post-graduate degree, spatial statistics.