After Flickr began using tag clouds in 2004 to display commonly used tags on its site (Viega & Wattenberg, 2008), it was quickly adopted by other platforms eg. del.icio.us (social bookmarking site) and Technorati (blog search & ranking). Tag clouds or word clouds are similar to a “weighted list”, whereby frequently used (aka popular) tags / text are displayed more prominently by using bigger font sizes. With the passage of time, the use of word clouds has gone beyond displaying popular tags on websites to the realm of information visualisation.

Flickr Tag Cloud, 2006
Image source: https://www.flickr.com/photos/cogdog/266534715/in/photostream/

Although word clouds have been in existence for more than 15 years, I did not have a need to use them until about a few months ago. I had a task of identifying frequently occurring terms used in 25 transcripts and had wondered how the work can be done with the least effort and at no cost. Word Clouds came to mind in a flash and I set out to find a tool that was free and easy to use.

I easily found recommendations on the internet and referred to:

Between the two posts, they introduced 11 tools, amongst which were Wordle, WordClouds.com, WordItOut.com. All but 3 were web-based or online tools.

Pro Word Cloud

Since the transcripts were saved as word documents, it would be more convenient to use tools that integrate with MS Word. Pro Word Cloud from Orpheus Technology Ltd came to my rescue. Pro Word Cloud is freely available for download from MS Store and it took less than 20 seconds to add and install.

To create a word cloud, I only need to:

  1. Enable the tool on MS Word
  2. Highlight the required text
  3. Click “Create Word Cloud”

In addition, Pro Word Cloud allows users to tweak:

  • Font / typeface: A choice of serif, sans serif and decorative fonts
  • Colours: It offers over 20 colour palettes
  • Layout: Mostly Horizontal, Vertical, Mostly Vertical, Half and Half, or Higgledy Piggledy
  • Letter Case : All Lower, All Upper, Intelligent or Preserve
  • Maximum number of words to display
  • Size of the cloud (in pixels)

Finally, there is also an option to remove common words (ie. stop words such as: and, of, the, to, was, …).

I can also copy and paste the word cloud in a word document or save the image in jpg, gif or png format.

Creating a word cloud using Pro Word Cloud was quick and almost effortless.  A question readers may have is, was it effective? Were the more frequently used terms also the important keywords? Before I answer this question, let’s take a look at the strengths and weaknesses of word cloud as a text analysis tool.

  • They are easy to create.
  • They are fast to produce.
  • We can see prominent terms at a quick glance.

As Word Clouds basically highlight relative frequency of words, the cons are:

  • Context is lost. We only get an indication about word frequency but do not have the context in which the words appear in the narrative.
  • Longer words could be more prominently displayed even though they occur less frequently.
  • We are unable to differentiate whether term was positive and negative: eg. like vs not like.
  • Same word but different meaning: eg. I like coffee vs We serve drinks like coffee and tea.
  • No feature to deal with text stemming:
    a) singular and plural nouns / verbs eg. choice vs choices or run vs runs.
    b) study, studying, studies. Hence, data cleaning may be required to group & edit similar terms, to deal with the negatives (eg. not-like), or to identify additional stop words.

Back to my application of word clouds, I found that the prominent terms were useful about only half of the time. There were also occasions when less prominent words were more relevant than the more frequently occurring ones. Thus, knowing what you intend to glean from a word cloud may help with the interpretation of the visualisation.

How to download Pro Word Cloud?
  1. Enable the Developer Tab
  2. Click Add-ins
  3. Go to Store
  4. Enter in search box: “word cloud”
  5. Click Add
How to deploy Pro Word Cloud?
  1. Open the required word document
  2. Enable the Developer tab
  3. Click Add-ins
  4. Select Pro Word Cloud
  5. Click Insert

For more information, refer to the following YouTube video created by The Tech Train (4:57 mins)

If you are interested to find out more about word clouds, look out for my next blog post on free web-based word cloud generators.

References

  1. Graham, S., Milligan, I., & Weingart, S. (2013). Basic text mining: word clouds their limitations, and moving beyond them in The Historian’s Macroscope – working title. Under contract with Imperial College Press. Open Draft Version, Autumn 2013. Retrieved from http://themacroscope.org
  2. S McKee. (2014, Feb 6). Presenting qualitative survey data with word clouds. [web log comment]. Retrieved from https://www.surveygizmo.com/resources/blog/qualitative-data-word-cloud/
  3. Viégas, F. B., & Wattenberg, M. (2008). Timelines tag clouds and the case for vernacular visualization. Interactions, 15(4), 49-52.

Be sure to check out other posts tagged Digital Scholarship, follow us @NTUsgLibrary and our hashtag #NTUsgLibraryDS on Twitter for more interesting tools and projects!