Steve Brewer, Director, Text Mining Solutions said: “In making best use of the power of text mining three key elements need to be combined: collecting, processing and visualising text data. The first two elements are addressed by a structured approach to searching and storing of data, but it is the third element – visualisation – which brings the results to life. By being awarded a Technology Strategy Board Innovation Voucher, we have been able to collaborate with the University of Sheffield to develop this new visualisation programme which we are now ready to take to market.”
Perhaps the most powerful way to communicate the task carried out by Prospector is to illustrate using a simple case study. Assume that we have retrieved from LinkedIn a set of data containing information on people, job categories and cities in the USA. “We can process and analyse this body of information using text-mining methods” says Steve Brewer, “then query the data set using Prospector. We can very quickly generate a visualisation graphic to show the relationship between job category and location – for example, there are a high number of jobs relating to politics in Washington.” Prospector allows complex relationships to be visualized in a simple way, and the records which go into the analysis to be retrieved.
A further example of visualising the correlation between topics in a large data set is shown in the diagram below. Here, a collection of over 150,000 scientific documents was analysed to show associations between 'drugs' and a number of other topics. The analysis highlights strong associations between drugs and herbs (constituents of many natural products), work (efficacy of drugs and workplace monitoring) and legislation. Prospector allows the data set behind such relationships to be mined very effectively.
Text mining offers potentially massive savings in time and effort for professionals working with large text data sets, and the ability to quickly query and visualise data is a further major step in confirming the benefits of text analytics. “This is the first visualisation tool available which can be used with records from any topic. Once TMS has collected, processed and mined a set of records, the end user can then interrogate the data at will using Prospector”, continues Steve. “We are genuinely excited by the introduction of Prospector as a tool to greatly increase the value of text mining to the end user.”