Tuesday, May 16, 2017
Azure Data Catalog – Tagging
Did you know that Microsoft has something called “Cognitive Services”? They are currently working on 25 Cognitive Services that they will be making available to you soon. These services allow you to build apps using Microsoft’s machine learning algorithms, using just a few lines of code. For now, they have three of these services that as of April 19, are either generally available or released as a preview. The first three Cognitive Services are:
1. Face API: This service allows you to detect and compare human faces, organize them into groups, and identify people you have tagged.
2. Content Moderator: This service allows you to automatically quarantine images, text and video for review prior to publication. The video moderator is currently only available as a preview, not for general use.
3. Custom Speech Service API. This service is currently only available as a preview. It allows you to customize Microsoft’s speech-to-text engine for your application.
Soon to be available for preview is a Handwriting detection and conversion API.
The full suite of Cognitive services can be found here: https://www.microsoft.com/cognitive-services
Tagging is the action of assigning a keyword or term to a piece of information. They are used to describe meta data items to allow them to be easily searched. They are used and determined by the creator. In ADC, the tags are located in the information pane. When you hover over a tag it gives information about the tag including who added it.
Note that anything can be added as a tag, even non-words. When adding a tag, if that tag already exists on that asset, ADC will automatically block the addition. This is true even when there is a change in case in the spelling of that word. ADC sees “WWWImporters” and “wwwimporters” as the same tag and will not include the second value.
Even without adding in duplicates it can be easy to see with the above example how the sort order can cause issues with Tags. The value 11 comes before 2 where as AAA comes before aa. Even thought I can not add aaa as we already have AAA. That does not remove the risk of ending up with a mess in your tags. The best way to limit this issue and others is to have a business glossary. In the next post we will look at setting up a business glossary in ADC