Music, Media & Metadata

I suppose on the commercial side we’re all aware of the growing awareness of using metadata to unlock stored value in monetizing music and media assets. One of the challenges in helping clients to better appreciate the value of big data is finding a way to help them to better understand it. I’m a storyteller as you know and a composer so let me take a crack at it.

 Metadata comes in two basic flavors: unstructured and structured. Structured data is neatly tied into hierarchical constructs that allow you to index and search; it’s not (generally) perfectly organized but there is at least a semblance of order and it is tagged with specific search parameters that make it possible to retrieve, categorize and prioritize. This allows for cross-functional search and makes is easier to catalogue information for multiple uses.

 Unstructured data is…well….just as it sounds. Not organized logically; languishing in the digital (or analogue) lofts of media and entertainment companies waiting for someone to discover it (content discovery) rationalize it and monetize it.


 I like to think of it this way: when I am composing a classical score every data point is defined and tagged across time and meter and pitch and timbre. It is notated in detail and can be recreated simply by deciphering the notation by putting it in front of a group of musicians (cryptanalysts).

 On the other hand there’s: Jazz! Jazz is improvised around a specific construct of harmony, melody and most of all, rhythm. In other words jazz has what we call “context”. There are basic parameters determined by the melody and harmony but the end result unfolds in free form in way that makes it very difficult if not impossible to recreate. Unstructured data; the information is there but you have to know where to look. Valid all the same but not until someone transcribes the work (which happens all the time) is it possible to access and aggregate the data. At that point it has been “tagged”, defined, notated and made accessible.

 Here’s another one: Theme parks generate unstructured data each and every day. Tens of thousands of data points (guests) making hundreds of thousands of decisions that result in recoverable data on rides, attractions, CCTV, retail, culinary and other points of access. Seldom is much of it collated or catalogued, aggregated or used. It’s just one massive dust storm of data that swirls about and then is gone. “poof”

 On the other hand there is baseball: A ball game is also a series of data points but here’s the difference: for over 100 years statisticians have recorded every detail of every pitch, hit and play making it possible to recreate the game and “play” it over. Brilliant. It is structured data that is used for everything from gaming to improving the team’s performance. It can be leveraged because it is already rationalized and there is a common language through which it can be shared.

 Play ball!

 © 2015 Minerd Music Works llc