Big Data Analytics: 2012 New Year's Predictions
By Philip Russom
Before January runs out, I thought I should tender a few prognostications for 2012. Sorry to be so late with this, but I have a demanding day job. Without further ado, here are a few trends, practices, and changes I feel we can expect in 2012.
Big data will get bigger. But, then, you knew that. Enough said.
The connection between big data and advanced analytics will get even stronger. My base assumption is that advanced analytics has become such an important priority for user organizations that it’s influencing most of what we do in business intelligence (BI), data warehousing (DW), and data management (DM). It even influences our attitudes toward big data. After all, the current frenzy – which will become more operationalized than ad hoc in 2012 – is to apply advanced analytic techniques to big data. In other words, don’t do one without the other, if you’re a BI professional.
From problem to opportunity. The survey for my recent TDWI report on Big Data Analytics shows that 70% of organizations already think of big data as an asset to be leveraged, largely through advanced analytics. In 2012, the other 30% will come around.
From hoarding to collecting. As a devotee of irony, I’m amused to see reality TV shows about collectibles and hoarding run back-to-back. Practices lauded in the former are abhorred in the latter, yet the line between collecting and hoarding is a thin one. Big data is a case in point. Many organizations have hoarded Web logs, RFID streams, and other big data sets for years. The same organizations are now turning the corner into collecting these with a dedicated purpose, namely analytics.
Advanced analytics will become as commonplace as OLAP. Okay, I admit that I’m exaggerating for dramatic effect. But, I have to say that big data alone has driven many organizations beyond OLAP into advanced forms of analytics, namely those based on mining, statistics, complex SQL, and natural language processing. This trend has been running for almost five years; there may be another five in it.
God is in the details. Or is the devil in the details? I guess it depends on what we’re talking about. With big data analytics, expect to see far more granular detail than ever before. For example, most 360-degree customer views today include hundreds of customer attributes. Big data can bump that up to thousands of attributes, which in turn provides greater detail and precision for customer-base segmentation and other customer analytics, both old and new.
Multi-structured data. Are you as sick of the “structured data versus unstructured data” comparison as I am? This tired construct doesn’t really work with big data, because it’s often a mix of structured, semi-structured, and unstructured data, plus gradations among these. I like the term “multi-structured data” (which I admit that I picked up from Teradata folks) because the term covers the whole range and it reminds us that big data is often a kind of mashup. To get full business value out of big data through analytics, more user organizations will invest in people skills and tools that span the full range of multi-structured data.
You will change your data warehouse architecture. At least, you will if you’re truly satisfying the requirements of big data analytics. Let’s be honest. Most EDWs are designed and optimized by their technical users for reporting, performance management, OLAP, and not much else. This is both a user design issue and a vendor platform issue. In recent years, I’ve seen tons of organizations rearchitect their EDWs (and sometimes swap platforms) to accommodate massive big data, multi-structured data, real-time big streams, and the demanding workloads of advanced analytics. This painful-but-necessary trend is long from over.
I’m stopping here because I’ve reached my target word count. And my growling stomach says it’s lunch time. But you get the idea. The business value of advanced analytics and the nuggets to be mined from big data have driven a lot of change recently, and will continue to do so throughout 2012.
SUGGESTED READING:
For a detailed discussion, see the TDWI Best Practices Report, titled
Big Data Analytics, which is available in a PDF file via a free download.
You can also replay my
TDWI Webinar, where I present the findings of the Big Data Analytics report.
For a discussion of similar issues, download the TDWI Checklist Report, titled
Hadoop: Revealing Its True Value for Business Intelligence.
And you can replay last month’s
TDWI Webinar, in which I led a panel of vendor representatives in a discussion of Hadoop and related technologies.
Philip Russom is the research director for data management at TDWI. You can reach him at prussom@tdwi.org or follow him as @prussom on Twitter.
Posted by Philip Russom, Ph.D. on January 23, 2012