WATCH: Unstructured data- Three main principles

Going Hybrid w/Deepak Subbarao: Unstructured Data

Add bookmark

Seth Adler

While enterprises know they have it, putting a definition to it is a little trickier. Deepak Subbarao breaks down what unstructured data is before jumping in on this episode of “Going Hybrid.” Next, he outlines three main principles to unstructured data management:

  1. Know your data. What story are your datapoints trying to tell you?
  2. Know the source of your data. Is it coming from emails, transcripts, videos, voice? The answer to this may lead to a change in strategy.
  3. Prepare your data. Simplify it, classify it, and label it.

All the fancy AI and ML technology is moot without workable data. Subbarao further breaks down each of his main points.

  1. In order to know your data, you must analyze it. In order to analyze it, you must first define a reasonable expected outcome. This works as a north star during data analysis and prevents the cut/paste mentality.
  2. Subbarao then offers a reality check. Resources are finite. Time, money, and processing power is finite. Making data work for you doesn’t mean throwing it against the wall to see what sticks. You must make it work for you by telling it what you want it to do. But first, you must know what you want it to do.
  3. A closing point is made on how to successfully merge unstructured data with structured data.

To hear Subbarao elaborate on the above, watch the full conversation below.