A brief tour of AR.Taxonomy

AR.Taxonomy is a desktop tool for interactively editing and testing taxonomies.

You can create a taxonomy from scratch, or import from a variety of sources including the curated taxonomies from WAND, Inc.

AR.Taxonomy is designed to be easy to use and dispenses with much of the jargon surrounding other semantic tools.  Despite it's simplicity, AR.Taxonomy is a very powerful tool to help you create and maintain data models for unstructured information.

The Taxonomy Pane

This is the taxonomy pane.

It uses the familiar tree view control found in many applications and utilities.  It is here that you can create, modify and delete taxonomy nodes, or categories.  Next to each category name is the number of documents that match the rule for this category.  This pane clearly shows the structure of the taxonomy tree and allows you to arrange the categories in a meaningful way.   

Double-click on a category and it's rule is executed against the local full-text index.  This is how you test category rules for accuracy.

The Properties Pane

This is the properties pane.  Here, you enter the name of the taxonomy node, and optional short description and the query rule that defines which documents are classified under this node.

It uses a simple full-text query language that most web users are familiar with.  The bottom field shows the expanded query as the built-in search engine will see it.  Expert users can build complex and precise rules that use boolean operators, proximity search and other advance search techniques.

The Results Pane

This is the search results pane.  It uses the familiar grid table or spreadsheet style view.   Any column in the index can be displayed or hidden so that you are sure that you get the expected results.  The results pane shows which documents will be matched by the taxonomy rule.  

Use this view to refine your categories rules.  Since you see how they are performing in real time, you can easily tell when you need to add or remove terms to get the results you desire.

The Preview Pane

This is the preview pane.

There are five tabs on the preview pane: the web view tab, the properties tab, the matching categories tab, the plain text tab and the expanded query tab.

The web view tab displays the document from its original source.   This allows you to view the document in its native format from its native location with all formatting intact.  

The properties tab displays all search engine properties that were extracted from this document or passed-in to the indexer.  Often you can leverage this information to build more precise queries for metadata values that may not be available in the document text.

The matching categories tab displays which categories in the current taxonomy will match this document.  A document can be classified under multiple categories, so this is helpful to see where in the taxonomy this document fits.

The plain text tab displays just the text of the document with all formatting removed.   The current search terms are highlighted in the document so you can see how the query matches.

The explain query tab shows a very technical explanation of precisely why the current query matched this particular document.  This tab is mostly for true search engine geeks to see into the guts of the system.

Conclusion

As you can see, AR.Taxonomy is powerful and easy to use.   Not only can you edit the structure of taxonomies, but you can instantly see how the categorization rules will affect your data in real time.  When you are finished editing your taxonomy, you can publish it to AR.TaxonomyServer where your rules add structure to unstructured data.

Structure gives context, and context makes it easier for end-users to find the information they need.