I've got scale on my mind! While speeding down the rails from Brussels to Paris on the TGV (the sleek, high-speed train), the scale of speed is breathtaking. In previous generations, going from Brussels to Paris for a single-day meeting would have inevitably involved a plane, with check-ins, security, travel to the airport and inflexible schedules. Travel would have easily eaten half the day. Now, with the TGV, I’m door-to-door in two hours.
There’s a metaphor here for scalable, enterprise analytics. Typically in analytics, there can be a great emphasis on speed. In the value case for the TGV, the speed of the train naturally plays a huge role. But so do other factors:
- Automation of check-in.
- Seamless handoffs from one train to another.
- Convenience and flexibility with trains every half hour.
- Direct travel from city-center to city-center.
At these levels, airplanes cannot scale, where high-speed trains can.
In my mind, scale means meeting growing demands with a resource pool that does not grow equivalently. Scale includes speed, but also a lot more. At SAS, we provide analytics in a number of different ways that all contribute to the scalability of enterprise analytics.
In enterprise analytics initiatives, the resource bottleneck can sometimes be the number of available data scientists. As the analytic appetite within an organization grows, we need systems and processes that will allow data scientists to spend more time innovating and less time maintaining operational processes: That’s offering decisions at scale!
Analytics drive better decisions in an increasingly competitive market. For years, analytics have been core to how organizations find insights to drive strategy. Which markets to focus on? Which products? And which customers?
Unfortunately, such analytically-derived insights have been difficult to consistently push down to the level where they can have the most impact: The operational touchpoints with customers, suppliers and within the organization. This is all changing. With an exciting range of new operational analytics solutions, such as Factory Miner and Decision Manager, SAS can enable “Decisions at Scale.” Whether the challenge is the scale of the data, the number of decision points, the complexity of the required analytics, or simply being able to do more analytics with the same number of analysts, SAS is upping the ante and bringing unprecedented ease, sophistication and automation to operational analytics.
Scale also means making certain evaluation tasks available to wider profile of business users. That’s approachable analytics!
With growing volumes of data, opening up powerful analytics to an army of knowledge-hungry business analysts can help uncover pockets of potential untapped by the data scientist alone. SAS offerings such as Visual Analytics and Visual Statistics allow for interactive data exploration and prototyping of powerful predictive models. Additionally, by making these powerful analytics easy to execute and interpret, more analysts can better understand the underlying methods to put them to better use.
Scale also means data preparation happens in a robust, automated way that’s uniquely relevant for the kinds of innovative analytics that motivate data scientists and which organizations need to drive competitive edge: That’s data preparation for analytics. Innovative new SAS offerings like Data Loader for Hadoop are making robust data preparation easier, and more automatic than ever.
Scale also means going closer to the edge of where data is actually created, then using analytics as an intelligent filter to capture only the relevant data. In an exploding sea of big data (Internet of things, sensors, social media, transactions, flash trading, etc.) Event Stream Processing becomes a vital part of any organizations’ big data strategy. That’s streaming analytics!
Further, scale also implies expanding the data types the organization can exploit. Documents and web pages are unstructured data, but sometimes contain vital information. Storing such textual data in a traditional database or appliance is neither efficient nor practical. SAS Solutions like Contextual Analysis combine the right storage medium with the relevant analytic techniques: That’s text analytics!
Last but certainly not least, scale also means driving unprecedented innovation. It means expanding the analytic and algorithmic possibilities available to data scientists. And then having the equally modern abilities to quickly and intuitively disseminate the found insights. That’s machine learning.
All of these analytics use cases are about driving the scale of analytics within the organization. They are about driving analytics in action.
For further discussion, have a look at the Machine Learning WebEx that Patrick Hall and I recently produced with O'Reilly.