www.mimeria.com
Big Data == Lean Data
Agila Sverige, 2018-05-30
Lars Albertsson
www.mapflat.com, www.mimeria.com
1
www.mimeria.com
Service-oriented architectures
● Services own data
● Heterogeneous coupling
2
Service Service Service
App App App
Poll
Aggregate
logs
NFS
Hourly dump
Data
warehouse
ETL
Queue
Queue
NFS
scp
DB
HTTP
DB DBDB
www.mimeria.com
● Teams own services
Service-oriented organisations
3
www.mimeria.com
● Need data from teams
○ willing?
○ backlog?
○ collected?
○ useful?
○ extraction?
○ data governance?
○ history?
Data-centric innovation
4
www.mimeria.com
● Need data from teams
○ willing?
○ backlog?
○ collected?
○ useful?
○ extraction?
○ data governance?
○ history?
Innovation value stream mapping
5
www.mimeria.com
Enter Big Data
● What is Big Data?
6
AI magic
Clusters
Weird technology
?
Spoiled developers
www.mimeria.com
A collaboration paradigm
7
Stream storage
Data lake
Data
democratised
www.mimeria.com
Onboard driven by use case
8
Data lake
www.mimeria.com
Data platform == collaboration platform
9
Data lake
www.mimeria.com
Balance of success
10
Data lake
Balance planning & architecture
● Homogeneity
● Governance
● Coordination
with business value driven activities
www.mimeria.com
Coupling by design
11
Data lake
● Coordination >> autonomy
● Homogeneity >> heterogeneity
www.mimeria.com
Data agility
12
Data lake
● Siloed: 6+ months
● Autonomous: 1 month
● Coordinated: days
∆
∆
Latency?
www.mimeria.com
A journey of learning
13
Data lake
● End-to-end == feedback, value
● Scale == cost
● All data now == waterfall
www.mimeria.com
End-to-end >> scale
14
AI magic
Clusters
Weird technology
?
Workflow orchestration
Proof of value
!
www.mimeria.com
Can I have AI now?
15
● Crawl, walk, run
AI
Deep learning
A/B testing
Machine learning
Analytics
Segments
Curation
Anomaly detection
Data infrastructure
Pipelines
Instrumentation
Data collection
Credits: “The data science hierarchy of needs”,
Monica Rogati
www.mimeria.com
A journey of many years
16
● Simple == max value
○ Reporting
○ Forecasting, risk
○ User notification
● AI first == waterfall
AI
Deep learning
A/B testing
Machine learning
Analytics
Segments
Curation
Anomaly detection
Data infrastructure
Pipelines
Instrumentation
Data collection
Value Effort
Credits: “The data science hierarchy of needs”,
Monica Rogati
www.mimeria.com
Who's talking?
17
Lars Albertsson
Mapflat - independent consultant
Mimeria - data-value-as-a-service
AI
Deep learning
A/B testing
Machine learning
Analytics
Segments
Curation
Anomaly detection
Data infrastructure
Pipelines
Instrumentation
Data collection
Credits: “The data science hierarchy of needs”,
Monica Rogati

Big data == lean data