4 September 2018

We caught up with Taxonomy Boot Camp London Day Two Keynote Tom Reamy during the summer to find out more about what he does and what he'll be talking about at the conference next month ...

Tell us about your current role and what you do
For the last about 13 years, I have headed up a consulting company, KAPS Group.  The company started doing taxonomy development and consulting and while we continue in those areas, we have shifted most of our focus to text analytics which, of course, utilises taxonomies and other knowledge structures.  To my mind, it makes no sense to do taxonomies without text analytics and despite what some technophiles claim, it makes no sense to do text analytics without taxonomies. 

In 2015, Information Today published my book, Deep Text, which is a comprehensive look at the entire field of text analytics based on my various projects for organisations big and small, commercial and government, and in keeping with the location of this conference, Taxonomy Boot Camp London, we even did one taxonomy project for the British Parliament.

Why are taxonomies more important than ever?
That is easy – more and more information and especially the information in unstructured content.  The increase in the value and importance of taxonomies is also fed by the increase in the use of taxonomies and good old competition – if your competitors are using taxonomies to build more advanced applications, you better jump on the bandwagon or be left behind.

What makes a good taxonomist?
A deep theoretical knowledge of language and a deep practical knowledge of the role of information in modern organisations.  Without the first, you get taxonomies that are shallow and superficial.  Without the second, you get overly abstract and useless taxonomies.

Oh, and a lot of hard work and a mind-set that revels in wading deep into details. 

How might the development of new technologies such as AI affect how we do our work as information professionals?
As far as AI and taxonomy, the value proposition is almost all in the opposite direction – the value that taxonomy brings to AI.  Modern AI is dominated by machine learning and deep learning neural nets which are great technologies for pattern and perception-based applications but not very useful for language and conception.  If you want a self-driving car you might have to wait a couple of years.  If you want a self-organising taxonomy you might have to wait a few decades.

On the other hand, the technology that continues to have a huge impact on taxonomies is text analytics.  First, text analytics enables organizations to “Mind the Gap” between taxonomies and documents with the use of auto-categorization to enable hybrid solutions that combine machine tagging with human tagging.  This dramatically improves the quality and consistency of tagging while lowering the cost.

Second, text analytics tools are a great way to help build taxonomies.  There are many ways that text analytics can help build taxonomies, but two of the most important are by analysing document sets for frequent and co-occurring terms and the clustering capability can also be used to explore related terms to flesh out the taxonomy.

What's the most exciting change you've seen in the industry in the last few years?
Exciting and taxonomy?  Not something you usually think of together and probably the most exciting change is in the related field of text analytics.  However, there have been a number of trends that are pretty exciting for those of us who take language seriously.  The growing acceptance of the importance of taxonomies.  The development of different kinds of taxonomies such as emotion taxonomies and action taxonomies. The growing number and variety of applications that can utilise taxonomies.

What will you be speaking about at Taxonomy Boot Camp London?
I will be speaking about how to make the business case for taxonomy.  Even though taxonomies continue to grow in popularity and status, there are still many organisations whose C-Level officers remain unconvinced of their value.  This is particularly true for technical companies who often believe that the software can do it all without the need for humans or such intellectual structures as taxonomies or ontologies. 

So, how do you make the case for taxonomies to all the myriad audiences?  Start with no one size fits all, make the usual ROI calculations, add in stories of successes and more importantly the spectacular failures of the cost of not having taxonomies, and enlist champions within the organisation.  But finally, the best argument is that organising the intellectual property of an organisation is as essential as organising the company into departments and roles.   It makes no sense not to do it.

The day before the conference, on Monday 15 October, I will also be giving a three hour workshop on text analyticswhat it is, the business case for it, the full range of applications that can be built with it, how to do text analytics development – categorisation, sentiment analysis, and data extraction, and finally how to get started  in text analytics or how to take your current under-performing initiative to the next level.

More News