a bit about architecture 1. “information architecture is a high level or general view of something...
TRANSCRIPT
2
“Information Architecture is a high level or general view of something that conveys an overall understanding of its various components and how those components interrelate.”
John Hobbs
7
Why is Architecture Important
• Achieve intended goals• Control weaknesses and threats• Specify and manage policies and
mechanics for delivering strategic goals• Defines infrastructure requirements • Minimize vendor dependence and cost• Drives effective governance
8
Architecture vs Infrastructure
“Infrastructure are the technologies required to support all the information systems activities taking place across the organization. The infrastructure will serve the users within the business much the same way a road and rail networks serve transport users.”
Source: Gunton T. “Building a Framework for Corporate Information Handling”, Prentice Hall, 1989.
Building a Healthcare Analytics Architecture
orWhat Would Dr. Snow Do?
a healthcare analytics thought experiment
10
What is a Thought Experiment?
A thought experiment or Gedankenexperiment(from German) considers some hypothesis, theory,[1
or principle for the purpose of thinking through itsconsequences. Given the structure of the experiment,it may or may not be possible to actually perform it,and if it can be performed, there need be no intentionof any kind to actually perform the experiment inquestion. The common goal of a thought experimentis to explore the potential consequences of the principle in question.
- Widipedia -
Our ThoughtExperiment Today
The Setting:Cholera Outbreak in London, 1854
Dr. Snow’s Study of the Epidemic and his Intervention
C
Can we conceive an analytic architecture capable of reproducing Dr. Snow’s results?
C
C
12
Dr. Snow andthe London Cholera Outbreak of 1854
Cholera – a disease of urban populationdensity (First Cholera in London – 1832)
Sudden outbreak in London’s Soho District, August 1854
Can kill within hours of onset
Extreme fluid loss
Blue skin tint in later stages
No germ theory
Miasma prevailing theory of cause
13
Our Protagonists
Noted anesthesiologistPrevious study of choleraSoho residentPublished soiled water theoryTheories shunned by community
Dr. John Snow Henry Whitehead
Assistant curate at St. Luke’sVery familiar with local custom and cultureOriginally believed ‘miasma’ theory
Our Antagonists
The real cause of Cholera
V. Cholereaa bacterium
William Farr
‘Miasma’ theory of disease predominatesSupported medically and politically“All Smell is Disease”Many ancillary miasma theoriesChloride of lime on streets
14
What Can We SayAbout Dr. Snow’s Data?
The London Census(The General Registry)
NameBirthDeath Record Name Gender Address Cause of DeathMarriagesProfessionAddress
Dr. Snow’s Data
NameDate of Fatal Cholera Attack(added from his interviews)Date of death(from the General Registry)Age (estimate)AddressAnecdotal information about ‘consumed|water source’. Did not carry out comprehensive or thorough survey
Whitehead’s Data
NameAgeAddress (assumed not explicitly stated) Position of the rooms occupiedSanitary arrangements, Consumed water with respect tothe Broad Street pump, and the hour of onset of the fatal attack.
Disparate SystemsNo initial integrationNo data integrity checkNo identifying index numberManually CollectedText based
15
Whitehead’s Corroboration
Located ‘Index’ patient(Infant)
Isolated probable cause of contamination(Soiled nappies thrown in nearby cesspit)
Caused cesspit inspection(Brick deterioration causing leak into Broad Street Well)
Abandoned disease theory of Miasma
Critical cultural and social knowledge keyleading to intervention
17
How does Dr. Snow take his data andchallenge a medical theory long entrenched in the medical, social, and politicalinstitutions of his world???
19
Snow’s Ghost Map Version 1:Not Good Enough for the Miasmists
‘Stacked’ deaths for emphasis
Broad Street pumpcommon water source
‘Look for life where there should be death. Look for death where there should be life.
The aunt and her niece
The workhouse
The brewery
20
Snow’s Ghost Map Version 2:The Voronoi Diagram Points inside Snow’s
diagram are closer tothe Broad Street pumpthan any other pump.
NOTE: Voronoi diagrams are named after Ukrainian mathematician Georgy Fedosievych Voronyi (or Voronoy) who defined and studied the general n-dimensional case in 1908
21
The Intervention
Whitehead discovers index patient’s father contracted cholera at the time of pump handle removal.
22
Could We Help Dr. Snow Today• Source data captured
all deaths logged by date• Define business rules
select only Cholera victimsreconcile patient identity and address
• Combine data from disparate data sourcesmashup – London City Map and Logged Cholera Deaths
• Cleanse Data. Explain data anomalies/outliersconversations. visitation
• Develop effective communication of resultsgraphic (not text) presentation
• Develop interventionremove pump handle
• Track post intervention resultslog of daily Cholera deaths
Infrastructure or Architecture ????
25
Source Systems
Load Original Data
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
1 Create Relational Database Warehouse Clean up, De-Dup etc Local Data Sources No Transformation – Preserve Original Data
26
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
Create Data Staging Layer Add Ancillary Data from Trusted Sources
DW Data Staging Area
27
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
Create Transforms and Business Rules Standardized Data Definitions Standardize Transformation Algorithms Group Like Entities (e.g. Master Person Index, Locations, Families, etc.
DW Data Staging Area
Business Rules
28
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
Create a Conformed Data Model Data Standards Applied to Original Data Transform Algorithms Applied Entities (people, families, locations) grouped correctly.
DW Data Staging Area
Business Rules
Create Conformed Data Model
29
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
.
DW Data Staging Area
Business Rules
Create Conformed Data Model
Analytic tools
Logical GroupingsLogical
GroupingsLogical GroupingsLogical
Groupings
Create Analytics Layer Analysis Tools Data Groupings (e.g. Cubes)
OutputReportsDashboardsScreensAlertsMaps
30
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
.
DW Data Staging Area
Business Rules
Create Conformed Data Model
Analytic tools
Logical GroupingsLogical
GroupingsLogical GroupingsLogical
Groupings
Create Analytics Layer Analysis Tools Data Groupings (e.g. Cubes)
OutputReportsDashboardsScreensAlertsMaps
31
Source Systems
DW – Clean, Reconcile, Combine, De-dup, standardize, transform
The 1850 Census
Dr. Snow’s Death Record
Whitehead’sInterviews
.
DW Data Staging Area
Business Rules
Create Conformed Data Model
Analytic tools
Logical GroupingsLogical
GroupingsLogical GroupingsLogical
Groupings
Create Data Governance Define Workflow Maintain Data Dictionary Insure Calculation Integrity
OutputReportsDashboardsScreensAlertsMaps
G O
V E
R N
A N
C E
32
Let’s Review Our Architecture
• Achieve intended goals• Control weaknesses and threats• Specify and manage policies and
mechanics for delivering strategic goals• Defines infrastructure requirements • Minimize vendor dependence and cost• Drives effective governance
33
What Problem Will You Solve Today?
When will you make something cool?
When will you make something
useful?
Young Geek Old Geek