1. Digging Up the Gold in your
Godowns
- A Big-Data primer for the busy Managers
2. What is Big-Data Anyway?
Well, conceptually a data set so large that your regular databases
can’t handle it is referred to as ‘Big-Data’.
Just how large is large ? Typically more than a few Petabytes.
( 1 Petabytes (PB) == 1000 Terabytes == 1000,000 GigaBytes == I billion
Megabytes)
A dataset of couple of Petabytes size currently is referred to as
‘Big Data’
And you will not necessarily need a big Infrastructure to tackle it!
3. What Big Data Infrastructure?
You can buy big machines…become bankrupt on the way and file
for Chapter 11!
else you buy commodity machines and put up a ‘Big-Data’
infra……
Open-source Non-open Source
Apache Hadoop,
HDFS , Hive, Pig, Flume, Splunk etc .
Wrappers like
Scoobi, Scalding, Apache Scrunch
can boost your productivity
Commercially supported
platforms like
• Cloudera,
• Horton Works,
• MapR
4. Where do you begin?
Have a look at the Big-Grid of Big-Data! below… What you want to do
may be within these four areas depending on your org’s domain….
Social Analytics Decision Support
Performance Optimization
(conventional BI/PA)
Data Exploration
Standardized Measurement Adhoc/Hypothetical/Scenario-driven
Transactional/Structured
Data
Non-Transactional/
Un-structured Data
5. In the Trenches- Performance Optimization
Done usually with business intelligence tools that provide
Canned queries and multi-dimensional analytics that provide
answers in real-time on structured/cleansed/formatted
data.
Standardized dashboards to view data that help in everyday
decision making for managers on the go.
6. You need to quantify abstract terms like
• Exposure
• Customer Engagement
• Virality
• Digital Presence
• Influencers
• Followers
• Customer Relationship Nodes/Graphs
And derive cognizant metrics specific to your organization.
In the Trenches- - Social Analytics
7. This has to do with using your current data to correlate significant content
/patterns and extrapolate findings to improve customer cross-sell,
engagements, conversion-rates, deliver targeted advertisements all with the
purpose of increasing the revenues per customer or adding more customers.
Also it can be employed to detect real-time fraud such as credit-card frauds,
system-attacks etc.
In the Trenches- - Data Exploration
8. This involves nerdy data scientists and statisticians crunching piles of un-
structured data running Adhoc queries , scenarios-based data-jungle
escapades ,integrate real-time feeds into the data that may involve social
processing like sentiment analysis, opinion scores and integrate the overall
feedback into the organization’s day-to-day strategies and decisions. This
involves utilization of lot of human cognition & perceptive faculties.
In the Trenches- - Decision Support
10. Big Data Capability – Concept to Execution
Setting up Big Data
Policies/Governance
Model
Setting up
Infra/Capa
bilities
Devising Big
Data
Strategies
Setting up Pilot
Projects & Begin
learning on the
way