Narrative to accompany the slides for our talk

These are our notes for preparing our slides for our CAA2018 presentation in the session “R as an archaeological tool: current state and directions” (https://github.com/MartinHinz/isaak_conf_sessions/tree/master/caa_2018_tuebingen)

Slide: Title slide

Reproducible research in archaeology using R & rrtools: In this talk, we will show how archaeologists are changing their data analysis practices to make their research more reproducible and open. We will introduce the rrtools package, which is our specific contribution to help with the challenge of making reproducible research in archaeology easier.

Acknowledgements

Our project comes from the group who took part in the 2017 Summer School on Reproducible Research in Landscape Archaeology at the Freie Universität Berlin (17-21 July), funded and jointly organized by Excellence Cluster 264 Topoi, the Collaborative Research Center 1266, and ISAAKiel.

Slide: Motivation & Outline

Motivation

Our motivation for this topic arises from a profound change that we have been observing in how researchers analyse their data.

Increasingly, we are seeing researchers publishing research that uses a programming language to calculate their models. Similarly, we see journal articles accompanied by scripts of programming language so that readers can study the details of the data analysis methods.

This is an important change away from using mouse-driven, point-and-click software for data analysis. When archaeologists use software such as SPSS and Past (as you see in our screenshots here), it is very difficult for them to share the details of their calculations. When readers cannot see the details of an analysis, it is difficult for them to be confident in the results described in an article. Also, for other researchers it may be near impossible to implement the described methods in their own work as they lack the details on how to come to the desired conclusions. Thus, using scripting language offers the chance of accelerating archaeological research by providing a way to retracing and checking published analyses as well as explaining methods more efficiently. These are the main advantages of creating reproducible research, which we would like to encourage.

Outline

In this presentation we will briefly review some data showing change in the use of programming languages in the sciences generally, and in archaeology in particular. We focus on the R programming language, because this is by far the most widely used language in archaeology.

Secondly, we will review how R is used by archaeologists, to show the potential that using a programming language has for opening and accelerating archaeological research.

Finally, we will briefly introduce our R package, rrtools, which is designed to simplify and standardise the process of using R for archaeology, and making research more open and reproducible.

Slide: Current uses of R [signpost slide only]

So, let’s start with the current uses of R.

Slide: Articles in all sciences citing R

This plot shows how data analysis is changing in the sciences more broadly towards the use of the R programming language. We searched the Web of Science, produced by Clarivate Analytics, for citations to the R program over the last ten years, and filtered the results to the top ten journals with the highest number of citations to this software. The purpose of this filtering step was to find areas of research were R is generally known, so we can see how its use is changing over time. Note that many articles that use R do not cite the R program, so these numbers underestimate the true use of R in science.

We can make two major observations:

First, the areas of science that have seen big growth in the use of R are ecology, evolution and related biological sciences. This is especially significant for us archaeologists, as ecology and biology are subjects from which we quite often adapt methods for analysis, as for example in point-pattern-analysis and statistical modelling. We also see increases in multi-disciplinary journals such as PeerJ and PLOS One.

Second, the scale of the increases is quite impressive: Starting only 2012 some journals now have 20% of articles citing R.

So, we can see this is a substantial change in the way researchers are working in many fields. What about archaeology? Do we also see a change like this?

Slide: Archaeology articles citing R