This document discusses best practices for big data analytics projects. It begins by defining big data and explaining that while gaining insights from large and diverse data sets is desirable, operationalizing big data analytics can be complex. It emphasizes understanding an organization's unique needs and challenges before selecting technologies. The document also explores how in-memory processing can help speed up analysis by reducing data transfer times, but only if the insights are integrated into decision-making processes.
2. Page 2 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
“Big data” has become one of the most talked-
about trends -- and yes, buzzwords -- within the business
intelligence (BI), analytics and data management markets. A
growing number of organizations are looking to BI and
analytics vendors to help them answer business questions in
big data environments. Unfortunately, gaining visibility into
pools of big data is easier said than done. And with vendors
marketing a wide variety of technology offerings aimed at
addressing the challenges of big data analytics projects,
businesses may be hard-pressed to identify the one that best
meets their needs.
‘Big Data’ Analytics Projects Easier Said Than Done—but
Doable
“Big data” has become one of the most talked-about trends -- and yes,
buzzwords -- within the business intelligence (BI), analytics and data
management markets. A growing number of organizations are looking to BI
and analytics vendors to help them answer business questions in big data
environments. Unfortunately, gaining visibility into pools of big data is easier
said than done. And with vendors marketing a wide variety of technology
offerings aimed at addressing the challenges of big data analytics projects,
businesses may be hard-pressed to identify the one that best meets their
needs.
So, what is big data -- really? A recent story by the IT publication eWeek
offered the following take on it, based partly on Gartner Inc.‟s definition of the
term: “Big data refers to the volume, variety and velocity of structured and
unstructured data pouring through networks into processors and storage
devices, along with the conversion of such data into business advice for
enterprises.”
That hits the mark in terms of data management and the analytics part of the
equation, but it misses the essential aspect of the business challenges
surrounding big data: complexity. For instance, big data installations often
involve information -- from social media networks, emails, sensors, Web
3. Page 3 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
activity logs and other data sources -- that doesn‟t fit easily into traditional
data warehouse systems.
And in many cases, all of that disparate data needs to be pulled together in
order to make sense of it on a broader level. That can have big implications
for business rules, table joins and other components of big data analytics
systems. The complexity of big data is what really makes it different from
more conventional data when it comes to storage and query management,
and it‟s the main reason why analytical database and data analytics software
vendors have had to step up their game to help companies deal with big
data.
Understanding big data is the first step in assessing your technology needs
and putting a big data analytics plan in place. The second is understanding
the market and the current trends that are affecting organizations looking to
derive business value, and competitive advantages, from increasingly large
and diverse data sets.
Big agendas for big data analytics projects
Many businesses have always had large data sets, of course. But now, more
and more companies are storing terabytes and terabytes of information, if not
petabytes. In addition, they‟re looking to analyze key data multiple times daily
or even in real time -- a change from traditional BI processes for examining
historical data on a weekly or monthly basis. And they want to process more
and more complex queries that involve a variety of different data sets. That
might include transaction data from enterprise resource planning and
customer relationship management systems, plus social media and
geospatial data, internal documents and other forms of information.
Increasingly, companies also want to give business users self-service BI
capabilities and make it easier for them to understand analytical findings.
All of that can play into a big data analytics strategy, and technology vendors
are addressing those needs in different ways. Many database and data
warehouse vendors are focusing on the ability to process large amounts of
complex data in a timely fashion. Some are using columnar data stores in an
effort to enable quicker query performance, or providing built-in query
4. Page 4 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
optimizers, or adding support for open source technologies such as Hadoop
and MapReduce.
In-memory analytics tools may help speed up the analysis process by
reducing the need to transfer data from disk drives, while data virtualization
software and other real-time data integration technologies can be used to
assemble information from disparate data sources on the fly. Ready-made
analytics applications are being tailored to vertical markets that routinely
have to deal with big data -- for instance, the telecommunications, financial
services and online gaming industries. Data visualization tools can simplify
the process of presenting the results of big data analytics queries to
corporate executives and business managers.
Organizations that fit into the categories described above in relation to their
data and analytics needs should begin by considering the following issues
and questions, among others, before creating an implementation plan and
finalizing their big data infrastructure choices:
• The required timeliness of data, as not all databases support real-time
data availability.
• The interrelatedness of data and the complexity of the business rules
that will be needed to link various data sources to get a broad view of
corporate performance, sales opportunities, customer behavior, risk
factors and other business metrics.
• The amount of historical data that needs to be included for analysis
purposes. If one data source contains only two years of information but
five are required, how will that be handled?
• Which technology vendors have experience with big data analytics in
your industry, and what is their track record?
• Who is responsible for the various data entities within an organization,
and how will those people be involved in the big data analytics initiative?
5. Page 5 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
Those considerations don‟t constitute in-depth requirements planning, but
they can help businesses get started on the road to deploying a big data
analytics system and identifying the technology that will best support it.
Match In-Memory Processing Speed to Big Data Analytics
Business Needs
Michael Minelli is vice president of sales and global alliances for the
information services division of MasterCard Advisors, the professional
services and data analytics arm of MasterCard International. He's also one of
the three co-authors of Big Data, Big Analytics: Emerging Business
Intelligence and Analytic Trends for Today's Businesses, a book that aims to
explain the big data phenomenon to both IT and business readers.
Minelli, who worked at software vendors Revolution Analytics and SAS
Institute Inc. before joining MasterCard, has been involved in big data
projects at organizations such as Time Inc., Cablevision, Foxwoods, Major
League Baseball, Standard & Poor's and Sony. In an interview with
SearchBusinessAnalytics.com, Minelli discussed big data analytics
applications and the role that in-memory analytics technology can play in
them. One of his pieces of advices: The faster performance supported by in-
memory processing won't provide the hoped-for business benefits unless the
analytical results are fed into real decision-making processes. Excerpts from
the interview follow.
What's the key message in your book?
Michael Minelli: The book's main theme is that big data analytics are a game
changer for the industry, whether you're in IT or the business. Big data is
going to make a big impact, and this is going to continue over time. The
message is to think about how to do things differently: "If we can do things
faster, then what does that mean for the business?" Big data allows us to
innovate and make decisions quickly while transforming the way we do
business intelligence. People have been talking about building one version of
the truth for a while -- that's the whole genesis of the data warehouse
movement. The name of the game was who could build the most valuable
6. Page 6 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
repository to make better decisions. What's changed is that it's not all about
what happens in your world, but what happens in other companies and even
in other industries. It's about going from having an insular mindset around
data to having an abundance mentality for leveraging data.
Can you give a couple of real-world examples of the opportunities for
taking advantage of big data analytics?
Minelli: Online up-selling and cross-selling on the fly before a customer's
attention fades. Mining blogs and customer service notes to perform
customer sentiment analysis, good and bad. Providing secure e-commerce
transactions with built-in anti-fraud controls. Deploying marketing automation
that delivers real ROI by enabling actionable insights into customer buying
habits. That's how competitive organizations are using big data analytics
today.
What specific role can in-memory analytics technology play in turning
big data into a competitive advantage?
Minelli: It's all about pure speed -- taking advantage of the hardware and
RAM capabilities that have become cheaper to create queries on the fly.
Doing so removes some of the barriers between IT and the business so that
there's more agility for people to do on-the-fly business intelligence and
predictive analytics and to move beyond traditional sampling techniques if
they don't have to wait 24 to 48 hours for results.
When is in-memory processing for big data analytics not the right fit or
more trouble than it's worth?
Minelli: It's the notion of fit to purpose. The main thrust is, do you really need
the additional speed and do you have the types of users and processes in
place so that if empowered with this information, they could really do
something with it to impact the business? Everyone talks about faster access
to data leading to better insights, but the other critical part is connecting that
insight to an actual operation. So it's not just viewing a report, but actually
using that report to make a decision on the fly to trigger an event like
addressing a fraudulent transaction or initiating a cross-sell opportunity.
Having faster analytics is great, but [the results] have to be able to make their
way into the decision process.
7. Page 7 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
So how can organizations get to the level where they truly can take
advantage of in-memory analytics in combination with big data?
Minelli: For starters, IT should work with the analysts to assess the low-
hanging fruit where there are some productivity gains [to be had]. A good
example is reducing a data mining process from 24 hours to a matter of
minutes so that a data scientist can be more responsive to the rapid changes
in today's businesses. The next step would be for the business to develop
some use cases where speed can make a difference and then give it a try.
The technology part isn't a no-brainer, but it's not quantum physics, either.
From my perspective, the major challenge is finding and hiring the right talent
and then managing that talent to achieve specific business objectives in a
reasonable time frame.
Beth Stackpole is a freelance writer who has been covering the intersection
of technology and business for more than 25 years.
8. Page 8 of 8 Sponsored by
Big Data Analytics Best Practices
Contents
„Big Data‟ Analytics
Projects Easier Said
Than Done—but
Doable
Match In-Memory
Processing Speed to
Big Data Analytics
Business Needs
Free resources for technology professionals
TechTarget publishes targeted technology media that address your need for
information and resources for researching products, developing strategy and
making cost-effective purchase decisions. Our network of technology-specific
Web sites gives you access to industry experts, independent content and
analysis and the Web‟s largest library of vendor-provided white papers,
webcasts, podcasts, videos, virtual trade shows, research reports and more
—drawing on the rich R&D resources of technology providers to address
market trends, challenges and solutions. Our live events and virtual seminars
give you access to vendor neutral, expert commentary and advice on the
issues and challenges you face daily. Our social community IT Knowledge
Exchange allows you to share real world information in real time with peers
and experts.
What makes TechTarget unique?
TechTarget is squarely focused on the enterprise IT space. Our team of
editors and network of industry experts provide the richest, most relevant
content to IT professionals and management. We leverage the immediacy of
the Web, the networking and face-to-face opportunities of events and virtual
events, and the ability to interact with peers—all to create compelling and
actionable information for enterprise IT professionals across all industries
and markets.
Related TechTarget Websites