Data in tabular form are ubiquitous in the social sciences. Cross-tabulations are used by criminologists, political scientists, archaeologists, sociologists, biologists, linguistics, to cite a few, as a convenient way of reporting and summarising data. However, tabular data can be thought of as being more than just summaries. Many patterns can be indeed embedded in such data, and bringing those patterns to the fore can provide many interesting insights.

Dr Gianmarco Alberti, resident academic lecturer at the Department of Criminology at the University of Malta, has authored a software to streamline the analysis of tabular data and to make the interpretation of the results easier. The software, named ‘CAinterprTools’, has been developed under the free R statistical programming language bearing in mind the anticipated needs of end-users who may be non-math oriented.

A small (yet complex) table where respondents’ ‘feeling of safety after dark’ (split between female and male) is tabulated against respondents’ ‘town size’ can be accessed online. The table is based on data from the International Crime Victim Survey that is freely available. In such a table, finding any obvious pattern is not an easy task. Dr Alberti’s R package allows, among other things, to carry out an analytical technique called correspondence analysis, whose perhaps most well-known visual output is portrayed in Figure 2 below or access the large version.

From the position of the points representing the row and column categories, we can see that as the town size increases (moving from left to right along the horizontal line) the feeling of safety after dark worsens. For both genders, the smallest town size goes hand in hand (i.e., is close in space on the chart) to the highest level of perceived safety. The lowest feeling of safety is associated to larger town sizes (500,000-1,000,000 and 1,000,000 plus). It is worth noting that women start feeling unsafe (‘bit unsafe’ category) in town of size 100,000-500,000 where men feel ‘fairly safe’ instead. Overall, it is apparent how the analysis can provide insights into the data and help revealing hidden patterns.

With an average of  500 downloads per month (source), and with 18 scholarly citations so far (source: Google Scholar), the software has been well received by users world-wide.