Thursday, November 18, 2010

Text Analysis Tools

We looked at text anlysis tools this week in class. I used Wordle and The Text Analysis Portal for Research (TAPOR) to analyse some speeches and statements I have come across this year. Wordle is a tool which allows you to enter text into the website and it creates a word cloud which highlights the most commonly used words in the text. As the website says, they create 'beautiful word clouds' and do give you a great indication of the most commonly appearing words.
           
TAPOR is a text analysis tool set up by the Canada Foundation for Innovation and the MacMaster University in Ontario. This website can analyse  the text of a HTML document from the internet or a TXT document saved on your computer. It lists unique words that appears in the document, the number of times they appear and you can see in what instances these words appeared.

When I started to use these tools I came across a problem in that most of the articles I hoped to use for text analysis came from the Irish Times Digital Archive, which displays articles in PDF format. As these articles are scanned it is impossible to copy and paste the text into the text analysis websites. After much frustration and failure, trying to use Evernote screen shots I eventually came across the image recognition site Online OCR. This website allows you to upload any scanned document whether it be a JPG image or a PDF scanned document and Online OCR will recognise any text on the documents enabling you to copy and paste it into TAPOR and Wordle.

I used these websites to analyse a minor dispute that arose between the Irish and British governments in early January 1978 over an interview Jack Lynch gave to RTÉ radio in which he stated his support for a power-sharing settlement in Northern Ireland and most controversially on a possible amnesty for Republican prisoners if a ceasefire were to be declared. Lynch issued a follow up statement days later after a barage of criticism from Unionist politicians and the Northern Ireland Secretary Roy Mason. The statement was published in the Irish Times  on January 10, 1978. Lynch angrily stated that at his previous meeting with Britsh Prime Minister James Callaghan  in September 1977 he had received a commitment from Callaghan of his determniation to reach a power-sharing settlement. I decided to compare the statement of Lynch in the Irish Times with the joint communiqué Callaghan and Lynch issued the previous September.
Below are the two Wordle word clouds.


Word Cloud for the joint communique of Jack Lynch and James Callaghan after their meeting, September 1977


Word Cloud for Jack Lynch's statement after comments on RTÉ radio, January 1978

Looking at this and the TAPOR results one can see  the words devolved and power-sharing feature prominently in Lynch’s speech. His attempt to play down the controversy over the Republican prisoner issue is seen is his mentioning of  the word amnesty only three times in the statement. It is also interesting to notice words that are not mentioned such as unity, a word Lynch used freely at the outbreak of the Troubles when passions were high and in opposition. It gives an indication of Lynch's tenative nature to calling for unity in his second term as Taoiseach, and his desire to avoid disrupingt the delicate situation in the North in the late 1970s.

In the earlier joint statement it isvisible Lynch had to yield considerably to Callaghan as the word power-sharing is not mentioned once! Words that reflect the pleasantries exchanged without any real progress dominated such as cooperation, community, talks and economic which  reflectied both government’s commitment to increased cross border economic cooperation. It is great that you can look beyond simple word counts in TAPOR, and look at the context in which they were used. In the case of the word cooperation I saw that in the joint speech this was used in terms of economic and security cooperation.
           
These text analysis tools are vital in examinig the content of political speeches and statements. They really reflect the  tensions and prominent issues which dominated in the example above. from what I have seen, Text Analysis tools can proove to be  invaluable in the analysis of political personalities and events.


No comments:

Post a Comment