Big Data is a catchy phrase. Unfortunately, it is often misused and misunderstood. Often, Hadoop and Big Data are used interchangeably; as if the Apache Hadoop family of projects are the only solutions for Big Data, or that that only use for these projects is from Big Data. Neither is true.
As an EDW/BI practitioner, I watched the Hadoop, or really, the Map/Reduce framework, be embraced and forced into being by software developers who were frustrated by Structured Query Language (SQL) and the need to create Entity-Relationship Diagrams (ERD) as data models or schæmas. They were equally unhappy with the various work-arounds to access Relational Database Management Systems from within their programs, such as Object Relational Models (ORMs) and Data Access Objects (DAOs). At first, I felt that these developers were simply lazy.
However, as I worked more with these so-called NoSQL technologies, it helped to clarify the dissatisfaction that I felt during the years I was leading EDW and BI projects. Thirty years ago, I worked in Aerospace System Engineering, developing methods and algorithms for risk assessment using Bayesian statistics. But, by 1996, I became involved in my first EDW project. Since then, the actual structure and functions associated with the data - defined by the data, became less important than fitting the data into an artificial structure imposed by business process models.
Don't get me wrong. Relational algebra, relational calculus and the DBMS technologies that came out of this mathematics, are all very useful. And, in the right hands, SQL is a very powerful language. ERDs provide a wonderful way to map data to business processes and to both transactional and analytic systems.
But… There is so much more that can be done with the data coming from traditional human-to-machine (H2M) interactions, but increasingly from human-to-human (H2H), machine-to-machine (M2M) and machine-to-human (M2H) exchanges. The interweaving of the flows of data from such disparate sources is what drives my research today.
- Gamification driving the adoption of smart meters for utilities
- Self-quantification use cases in the workplace
- Sustainability for increasing the bottom line
- Combing social media and sensor data for profitability
- Sensor analytics as an ecosystem
These, and over 70 other use cases that I'm cataloguing, come from the innovation surrounding hype of Big Data, and the Data Science movement. In a recent Quark, I've classified this innovation into 11 areas. A compete mindmap is linked from the initial mindmap shown below, and in the report.
The Quark covers the trends coming from these innovations, and develops the four keys required to bring valuable decision making processes into your organization from these innovations. It's entitled "Big Data: It's Not the Size, It's How You Use It". For such a simple report, it took over 8 months to develop. Mostly this delay was caused by the fast-paced evolution of the innovations. The executive summary from the Quark is linked from the title.
I hope that you find that information, as well as the mindmap, useful in incorporating inference, prediction, insight and performance with intuition for making better decisions.
