News from the Edge

A deeper look at taxonomy

Posted by Jon Campbell on Mar 17, 2016 11:05:56 AM

When I first started out as a business researcher the internet was in its infancy. Online research tools were basically paper directories that were scanned or typed into a digital format. The tactile feel of a thick paper directory was replaced by the blinking green cursor on a monochrome computer monitors. One of my first tasks was to pull up stories about Mergers in the Fertilizer industry. Sounds like an easy job, right? Well once I started tooling for the “Fertilizer” code my job got very difficult. There was no search function and I was forced to browse through the various branches in the tree. This is taxonomy speak for categories (sometimes called parents) and subcategories (sometimes called children) that resemble branches of a tree.

Getting back to my example, I thought I was on the right track when I saw Agriculture as a category. I reasoned that farmers used fertilizers and I would have the search nailed. Two hours, three cups of coffee and a few broken pencils later I finally found it, nestled under Chemicals – Organic & Specialty. There it was: Plant Nutrition/Fertliser, and with the British English spelling to boot.

I recognized two big problems, one technology related and one structural. The technological challenge was that there was no good way to search in the tool I was using. There was a team already working on that and I would have to wait for them to roll out the improved version of the product. The structural problem, however, was fixable. After this frustrating real-world experience I spent a lot of time thinking about the mono-hierarchical model of taxonomies. “Thou Shalt Not Have More Than One Parent” had been drilled into the heads of Librarians and Classifiers for so long that many never challenged this misguided notion.

One day a clear solution came to me while on vacation in Puerto Rico. I was kayaking through a mangrove swamp and I stopped to take in the scenery. There were so many leaves and so many branches. Try as I might, I could not trace a single leaf to a single tree. Branches split off, then rejoined each other before intermingling with another set of tree trunks.

That was it! The answer to my problem: Break the model! Build a new one that made sense.

Blog_25_Tree.jpgI created a proposal that showed a topic appearing in multiple places in the tree. Defined at the leaf level, that leaf could exist on a number of branches with no conflict and would be a rational part of its successive parents. I was able to convince others that this model made sense and would increase user productivity and adoption of electronic research databases.

Today, as I am responsible for the business information taxonomy at Acquire Media, I put that same philosophy into practice when we rebuilt our taxonomy a few years back. Recognizing that terms such as “Alcohol Licensing Law” could relate to both “Business Regulation” and “Government Regulation” means that we could have customers coming from differing vantage points looking for the same information. The single-hierarchy tree approach did not make sense then and certainly not now. Instead, we decided to treat our taxonomy like building blocks where users can use each parent (category) and child (sub-category) independently to find relevant information. Whether they are casting a wide net or looking for something very specific, every search is unique, and using our taxonomic terms to build that search gives users the ultimate control on their results set.

Depending on your job function, you may or may not think much about the taxonomy you are using to find information but they are pretty important. My career has taken me down the path of business news and information but there are so many different taxonomies and just as many approaches to utilize them. Take a few moments to think about what you use today. Is it relevant to my company? Does it offer the specificity I need? If the answer is yes, then you have chosen wisely. If not, take the time to do a little research now. It will pay off in the long run.

NewsEdge takes advantage of the multi-parent taxonomy – behind the scenes, where the user never has to worry about whether or not a story they’re looking for is categorized as an Organic Chemical or an Agricultural Chemical. Metatags in each article displayed in the NewsEdge interface show which categories an article comes under, making it easy to see how the multi-hierarchical model works.

Refresh your memory with last week's taxonomy introduction.

Try NewsEdge Today

photo credit: Hojas via photopin (license)

Topics: taxonomy, categorization, classification

Why read News from the Edge?

The NewsEdge Blog

NewsEdge, a service of Acquire Media, has been serving the information needs of busy professionals in corporate, finance and government for over 25 years.  We are experts in surfacing business relevant information through web, mobile and feed deliveries.  We specialize in content categorization and distribution to ensure users receive only the news they need, when they need it and how they need it.

Our blog aims to:

  • Discuss hot topics affecting the information industry
  • Offer our insights into new technologies - the good and the bad
  • Invite others to share viewpoints on how information is changing in real-world environments

Stay in Touch with Email Updates

Recent Posts