big mountain data workshop technical track

32
Data & Advanced Analytics Technical Track by ZIFF winners of the Big Mountain Data Competition ziff.io & @ziffio

Upload: david-b-gonzalez

Post on 27-Jan-2015

109 views

Category:

Technology


3 download

DESCRIPTION

Data & Advanced Analytics: Technical Track by ZIFF winners of the Big Mountain Data Competition ziff.io & @ziffio

TRANSCRIPT

Page 1: Big Mountain Data Workshop Technical track

Data & Advanced AnalyticsTechnical Track

byZIFF

winners of the Big Mountain Data Competitionziff.io & @ziffio

Page 2: Big Mountain Data Workshop Technical track
Page 3: Big Mountain Data Workshop Technical track

introductionswhat you hope to gain

Page 4: Big Mountain Data Workshop Technical track

Data Products

Data Aggregation

Analytics

Recommended Action

Activity

Page 5: Big Mountain Data Workshop Technical track

Objectives

● Easy to try n-scale nosql you can push to aws

● Use cases & Skullcandy● Relevant to you & your company

○ Problem types○ Solution types

Page 6: Big Mountain Data Workshop Technical track

Upfront Contract

YESBusiness value

Specific to your needs

Getting started and next stepsWell formed next-step

Collaboration

NOMe-too

Latest hotness

What the big guys are doingUnrealistic/overly complex

Long lectures

Page 7: Big Mountain Data Workshop Technical track

Agenda

1. Introductions2. Typical architectures3. Identify where to make an impact4. Identify requirements for a POC5. Wrap up

Page 8: Big Mountain Data Workshop Technical track

data collectors bulk store

batch process

livestore

apiservice ui

Datacenter AWSand/or“Analytic”

queues & oltp*SQSredis or couchmongodbrdbmsolapmongoHbaseThrift/Protobuf/AVROsockets style*messagepack basednettykafka*kinesisapps*elastic beanstalkec2/vm + load balance

emitters*messagepack

*s3ebs*HDFScassandracolumnar

on file systemM/R based (pig, hive)Graph based

off file systemanything language

diy jsonmongodb*BaaSespostgres

columnHbase/Impala*cassandra

graphcyphergremlin

searchelasticsearch

*stupid-simple-n-scale

cloud/dc appsec2/vm + load balance*elastic beanstalk

sql-ishPhoenixCassandra

graphGremlinCypher

searchelasticsearch

d3nvd3.js - simpled3.js - complexdc.js - dimensions

putting the long-A in OLAP

Page 9: Big Mountain Data Workshop Technical track

data stream process

bulk store

batch process

livestore

apiservice ui

Datacenter AWSand/orStream

just processstormspark + shark

machine learningstorm + tridentspark + mllib

Page 10: Big Mountain Data Workshop Technical track

impactalignment = buy-in

Page 11: Big Mountain Data Workshop Technical track
Page 12: Big Mountain Data Workshop Technical track

Product Innovation/Product Leadership

Customer Relationships/

Customer Intimacy

Infrastructure Mgmt/Operational Excellence

Economics Early market entry enables charging premium prices and acquiring large market share; speed is key

High cost of customer acquisition makes it imperative to gain large wallet share; economies of scope are key

High fixed costs make large volumes essential to achieve low unit costs; economies of scale are key

Competition Battle for talent; low barriers to entry; many small players thrive

Battle for scope; rapid consolidation; a few big players dominate

Battle for scale; rapid consolidation; a few big players dominate

Culture Employee centered; coddling the creative stars

Highly service oriented; customer-comes-first mentality

Cost focused; stresses standardization, predictability, and efficiency

Business

Business Model Generation

Page 13: Big Mountain Data Workshop Technical track

Product Customer Operations==MARKETING==

Who are your potential customers?

What do they want?Brand loyalty?What’s next?

e.gCustomer seg/profile

Market analysisSentiment

==SALES==

What motivates customers?

Which channels work best?

What else do they need?

e.g.Engagement

ChurnConversion

OffersChannel attribution

==OPS==

How long will X function?

How much product? waste?

Will Y be cheaper?

e.g.Yield optimization

Failure ratesFutures

Competing on analytics

Page 14: Big Mountain Data Workshop Technical track

Product focused problems

● Marketing○ Keywords○ Channel Attribution○ Rev/Cost per channel○ Targeting○ Profiling○ Markets (at large, PPC, etc.)

Page 15: Big Mountain Data Workshop Technical track

Customer focused problems

● Sales○ Sentiment○ Retention & Churn○ Profile○ Customer Support○ A/B & Multivariate○ Offers

Page 16: Big Mountain Data Workshop Technical track

Operations focused problems

● Ops○ Failure rate○ Yield optimization○ Pricing & Costing○ Time to ____○ Traveling Salesman○ Backpack

Page 17: Big Mountain Data Workshop Technical track

Key Partnerships

Key Activities

PLATFORM MGMT

MANAGING SERVICES

EXPANDING REACH

Value Propositions

TARGETED ADS

FREE SEARCH

MONETIZING CONTENT

Customer Relationships

Customer Segments

ADVERTISERS

WEB SURFERS

CONTENT CREATORS

Key Resources

SEARCH PLATFORM

Channels

Cost Structure

PLATFORM COSTS

Revenue Structure

KEYWORD AUCTIONS

FREE

Business Model Canvas: GoogleBusiness Model Generation

Page 19: Big Mountain Data Workshop Technical track

Gain, Pain, & Google

You

GainGainCreators

PainRelievers

Pain

JobsProducts & Services

FIND STUFF

TOO MANY THINGS

OFF TOPIC

SAVE TIME

PERSONALIZED

CAPACITY

TARGETING

SIMPLE UI

MAGIC

WEB SEARCH

Page 20: Big Mountain Data Workshop Technical track

Gain, Pain, & Customer/Revenue-side Analytics

You

GainGainCreators

PainRelievers

Pain

JobsProducts & Services

Page 21: Big Mountain Data Workshop Technical track

Gain, Pain, & Partner/Cost-side Analytics

Partner

GainGainCreators

PainRelievers

Pain

JobsProducts & Services

You

Page 22: Big Mountain Data Workshop Technical track
Page 23: Big Mountain Data Workshop Technical track

local to cloudM/R

AWS

Page 24: Big Mountain Data Workshop Technical track

exerciseidentify key components related to pain(s)

Page 25: Big Mountain Data Workshop Technical track

SkullcandyFocus: Get more customers with online channelKnow● Who● vs CompetitionBe able to● Target● Get in frontProfile● muses & personas

Page 26: Big Mountain Data Workshop Technical track

What must be true?

What must be true in order to detect/identify/catch ___________________?● know● be able to● profile (story)

Page 27: Big Mountain Data Workshop Technical track

… in order to get the info

● people● data points● resources

Page 28: Big Mountain Data Workshop Technical track

… in order to use it

● metric(s) or key results● tools● technology● expertise

Page 29: Big Mountain Data Workshop Technical track

Wrap up

HopeYou & your organization will incorporate

advanced analytics as part of your competitive strategy in your market.

Page 30: Big Mountain Data Workshop Technical track

Wrap up

BeliefData & Advanced Analytics can be key in better aligning your business with strategic objectives

and can be ideal for helping to measure key results in the course of meeting those

objectives.

Page 31: Big Mountain Data Workshop Technical track

Wrap up

Dare to dream● Innovation as accepted norm in your enterprise● Employees feel greater autonomy, mastery, purpose● Shared, well defined, measurable vision and objectives● Leaner, more productive organizations● Delight customers