SAMPLICS: ANALYZING THE POTENTIAL OF PYTHON FOR TEACHING SURVEY SAMPLING

 

Luis Castro-Martín¹*, Jorge Luis Rueda², Beatriz Cobo³

¹Dr Luis Castro-Martín, Andalusian School of Public Health, SPAIN, luiscastro193@ugr.es

²Mr. Jorge Luis Rueda, University of Granada, SPAIN, jorgerueda279@correo.ugr.es

³Dr. Beatriz Cobo, University of Granada, SPAIN, beacr@ugr.es

*Corresponding author

 

Abstract

Nowadays, teaching survey sampling is directly connected with the assisting software used for implementing the different techniques and applying the concepts in a practical scenario. Traditionally, the programming language R has been used for this purpose given its long history as the main language for statisticians. Alternatively, simpler programs such as SPSS are also common. However, there is a new tendency recently towards the programming language Python, of growing popularity. Even though it is intended as a general-purpose language, its success is being so significant, especially in relation to the field of data science, that it is being considered by many as the most interesting alternative for Statistics as well.

A good example of this new tendency is the publication of the 'Samplics' package. It translates advanced survey sampling concepts which until then were only available in specialized software such as packages for R or SPSS. This library and others such as Pandas, NumPy or SciPy represent an effective and convenient starting point for applying the theoretical concepts required in a real use case. They are well documented and backed by a consolidated community.

In this work we will detail the advantages that can be found when working with Python in general and the 'Samplics' package in particular for teaching practical techniques to students with no background experience. We will emphasize the ease-of-use, the clarity and legibility of the resulting code and the versatility of the open source tools associated with this ecosystem. Therefore, professors will find a new alternative with growing demand for their classes.

Keywords: statistics, teaching, open source, python