How the Democratization of Data Should Actually Work
June 18, 2014 Leave a comment
- Thanks in part to the cloud and pervasive mobility, vendors are looking to democratize data, putting advanced business intelligence tools into as many hands as possible.
- Placing advanced analytics tools into the hands of the average business not only highlights the importance of data expertise, but may also drive interest in what was a narrow field of endeavor.
This week, Microsoft announced plans for eventually rolling out an advanced analytics service called Azure Machine Learning. The aim of this service is to make predictive analytics accessible to a broad range of companies, not merely those equipped with an army of data scientists and a commensurately large software budget to match. That’s a pretty tall order and at first blush maybe not even be a good idea.
Do we really want to equip business users with a tool powerful enough to create or lose a competitive advantage depending upon whether or not that person has the appropriate understanding of predictive algorithms? Do we really want users with no deep knowledge of the principles of data modeling or perhaps even the R language itself playing with something like Azure Machine Learning? Of course not. Yet, the cloud demands it. Mobility demands it. The expectations of modern business demand it, at least in the desire to react to market opportunities rapidly by discovering future opportunities hidden away neatly within terabytes of historical data. The science – and it must be said, the art – of creating, testing, deploying and running a predictive business model is something best attempted by an expert.
Yet, that’s exactly where Microsoft may be onto something. While the vendor has indeed built a drag-and-drop interface capable of literally assembling predictive models, the intent is not to turn such work over to an intern. Well, if that intern just happens to be a data scientist, then yes, that’s exactly who Microsoft would like to target. By focusing on rapid modeling, test, deployment and execution all on the Azure services platform, Microsoft wants to speed up and greatly simplify the process itself — something the vendor hopes will appeal to the emerging data scientist. Moreover, by subsequently building a literal storefront for models on the Azure Store, Microsoft intends to establish an ecosystem of said professionals interested in making a business out of their expertise and experience.
This is where the democratization of data truly comes into play. Smaller companies, those without armies of data scientists and associated accoutrements, will be able to buy a well-tested and well-understood data model from the Azure Store and, with a little bit of consultation, put that model to work in a matter of days rather than months. Certainly, if giants such as Microsoft, Google, Apple and Salesforce.com have shown us anything, it is that developers drive markets. By applying the same model to the data sciences, perhaps we’ll see a similar evolution among data scientists.