Whilst it would be very nice if all our data collection and analysis was done by looking up into a big blue sky with giant fluffy white clouds, the clouds that this article will concentrate on are content clouds. A content cloud or word cloud is data visualisation summarising the contents of a document. These clouds are compiled in similar ways to tag clouds (those created from blogs) but use uploaded text instead. If you feel that you are not sure what a word cloud is from my description then it is more than likely that you have seen a content cloud on television or on the Internet. The media have used word clouds quite often, especially in political coverage for example when summarising a Prime Ministers speech and what the most frequently used word was.
Cidell in her recent article Content clouds as exploratory qualitative data analysis discusses how visualisation of data has become a way to ‘quickly display and summarise content’. Two case studies are given as examples of ways that content clouds can be used for exploratory data analysis, one being public meeting transcripts. As a research assistant a few years ago, finding new and interesting ways to show data was a challenge and so using focus group data I decided to use a word cloud (using Wordle software) to visualise some results. This example is similar to Cidell’s public meeting, where there are many voices to be heard, a content cloud can visualise the most frequent and possibly most important words.
However, Cidell does recognise in her article the various cautions that need to be kept in mind when using a method like this as exploratory data analysis. Important and frequent words can get pushed down the list by other words that may skew results and of course there are difficulties involved when analysing single words at a time. Nevertheless, the potential for content clouds to be useful as a form of data analysis is high and may contribute to wider geographical data analysis in time.
Cidell, J (2010) “Content clouds as exploratory qualitative data analysis“, Area, Volume 42, Issue 4 pages 514-523.