2012.12 - EDDI 2012 - Workshop

Create Your Own Information Systems
on the Basis of DDI-Lifecycle

4th Annual European DDI User Conference
(EDDI 2012)

03.12.2012

Thomas Bosch Matthäus Zloch
M.Sc. (TUM) M.Sc.
Ph.D. student Ph.D. student
boschthomas.blogspot.com
GESIS - Leibniz Institute for the Social Sciences GESIS - Leibniz Institute for the Social Sciences

Introduction of Attendees
• What is your name?
• What’s your organization?
• Why are you here?
• What are you mostly interested in?
• What do you expect from this tutorial?
• …

Appr. 30 seconds
2

Goals of This Presentation
• Introduction to DDI-Lifecycle
• Give a DDI Lifecycle overview
• Show you (basic) elements of the DDI-Lifecycle and the main structure
• How does Identification, Versioning, Maintenance work?
• DDI Discovery Vocabulary “disco”
• How to use the DDI Documentation
• Individual software project
• Introduce you to basics of software architecture design
• Show you how the software architecture of your project may look like
• Show you how a software project leverages DDI and the disco model
• Introduce you to a project template for using the disco model in the
back-end
3

Outline
Matthäus Thomas

• Missy – General Information • DDI-L Overview
• Requirements to Developers • Identification, Versioning, Maintenance
• Use-Cases in Missy • DDI-L Main Structures
• DDI Instance
• Study Unit

• Software Architecture • Conceptual Component
• Multitier • Logical Products
• MVC • Data Collection
• Missy Data Model • disco-model
• Extendable Data Model

4

Outline

• Persistence Layer • Physical Data Product
• Programming Interface • Physical Instance
• Types of physical data storages
• Archive
• Examples
• DDIProfile

• Outlook • DDI Serialization Examples
• DDI-XML
• DDI-RDF / disco-spec
• Relational-DB

5

Outline
Matthäus Thomas

• DDI Instance
Presentation
• Study Unit

• Extendable Data Model Business Layer

6

Outline

• Archive
• Examples
• DDIProfile Data Storage Access

• DDI-XML
• Relational-DB
Data Storage

7

Feel free to ask questions at any time!

8

What comes next?
• Before we start with the use-case Missy…

• Thomas, can you give us a top-level overview of DDI?
• You might also include how DDI treats identification, versioning and
maintenance…

9

What comes next?

• DDI Instance
Presentation
• Study Unit


10

using

Survey
Study
Instruments

measures made up of

about

Concepts Questions

Universes

with values of
Categories/
Codes,
Numbers
Questions

Variables

collect made up of

Responses
Data Files
resulting in

What comes next?

• DDI Instance
Presentation
• Study Unit


14

Element Types

IDs must be unique within maintainables
Not identified elements are contained by other element types
15

Identification

Agency, ID, Version

or

URN (= Agency + ID + Version)

16

Referencing Maintainables

17

Schemes

Very often, identifiable and versionable elements are maintained in parent schemes

18

Maintenance Rules
Maintenance Agency
• identified by a reserved code (based on its domain name; similar to it’s
website and e-mail)
• Registration of maintenance agency (e.g. GESIS)

Ownership
• Maintenance agencies own the objects they maintain
• Only they are allowed to change or version the objects

Reference
• Other organizations may reference external items in their own schemes,
but may not change those items
• You can make a copy which you change and maintain, but once you do
that, you own it!
19

Versioning Rules

20

Identifiable (DDI 3.1 XML)
<DataItem id=“AB347”>
…
</DataItem>

21

Versionable (DDI 3.1 XML)

<Variable id=“V1” version=“1.1.0” versionDate=“20012-11-12”>
<VersionResponsibility>
Thomas Bosch
</VersionResponsibility>
<VersionRationale>
Spelling Correction
</VersionRationale>
…
</Variable >
22

Maintainable (DDI 3.1 XML)
<VariableScheme
id=“VarSch01”
agency =“us.mpc”
version=“1.4.0”
versionDate=“2009-02-12”>
…
</VariableScheme>

23

URN (DDI 3.1)

urn:ddi:us.mpc:VariableScheme.
VarSch01.1.4.0:Variable.V1.1.1.0

24

URN (DDI 3.1)


25

URN (DDI 3.1)


Maintaining agency

26

URN (DDI 3.1)


Maintainable (element type, ID, version)

27

URN (DDI 3.1)


Identifiable (element type, ID, version number)

ID must be unique within maintainable

28

URN Resolution
• DNS-based resolution for DDI services
• URN can be resolved by DNS means
• http://registry.ddialliance.org/
• details and library
• list of available DNS services for this DDI agency

• URNs are mapped to local URLs
• XML Catalog can be used as simple resolution mechanism

29

Referencing (DDI 3.1)

(Agency, ID, Version) vs. URN
Often, these are inherited from a maintainable object
30

Referencing (DDI 3.1 + 3.2)

• Internal
• In same DDI Instance
• (agency, ID, version)
• Maintainable (agency, ID, version)
• Identifiable (ID), versionable (ID, version)

• External
• To other DDI Instances
• URN > (agency, ID, version)

31

Internal Reference (DDI 3.1)
<VariableReference isReference=“true” isExternal=“false”
lateBound=“false”>
<Scheme isReference=“true” isExternal=“false”
<ID>VarSch01</ID>
<IdenftifyingAgency>us.mpc</IdentifyingAgency>
<Version>1.4.0</Version>
</Scheme>
<ID>V1</ID>
<IdenftifyingAgency>us.mpc</IdentifyingAgency>
<Version>1.1.0</Version>
</VariableReference> 32

External Reference (DDI 3.1)

<VariableReference isReference=“true” isExternal=“true”

<urn>
urn:ddi:us.mpc:VariableScheme.VarSch01.1.4.0:Variable.
V1.1.1.0
</urn>

</VariableReference>
33

Identifiers (DDI 3.2)

• Why new IDs?
• better support of persistent identification (less semantics, context
information)
• URN or (agency, ID, version)
• Version changes?
• Administrative metadata  No
• Payload metadata  yes
• No inheritance
• Why? (implementation problems, …)
• Deprecated vs. Canonical IDs
• Deprecated (no version of maintainables, colons as separators)
• Mechanism for roundtrip available between DDI 3.2 canonical ID, DDI 3.2
deprecated ID, DDI 3.1 ID
34

Canonical URN (DDI 3.2)


No maintainables

35


urn:ddi:us.mpc:Variable.V1.1.1.0

No maintainables

36



Agency, ID, version (Identifiable, Versionable)

37



ID unique in agency (not maintainable)

38



No element types

39


urn:ddi:us.mpc:V1.1.1.0

No element types

40



Colon is separator
41


urn:ddi:us.mpc:Variable.V1:1.1.0

Colon is separator
42


urn:ddi:us.mpc:V1:1.1.0

No maintainables
Agency, ID, version (Identifiable, Versionable)
ID unique in agency (not maintainable)
No element types
Colon is separator
43

What comes next?
• This was the DDI Overview with identification, versioning, and
maintenance

• But, what is the use-case actually?
• What do you want to use DDI for?

44

What comes next?

• DDI Instance
Presentation
• Study Unit


45

General Information About Missy
• Missy – Microdata Information System

• Largest household survey in Europe
• Provides detailed information about individual datasets
• Currently consists of “microcensus" survey, which is
comprised of statistics about, e.g.
• general population in Germany
• situation about employment market
• occupation, professional education, income, legal insurance, etc.

46

General Information About Missy
• May be split in two parts
• Missy Web (end-user front-end part)
• Missy Editor for documentation (back-end part)

• Consists of approx. 500 Variables & Questions
• Captures 25 years, since 1973

47

Missy 3
• That’s what it’s all about! The “next generation Missy”

• Integration of further surveys, e.g. EU-SILC, EU-LFS, …
• Implementation of Missy Editor as a Web-Application

48

Use-Cases in Missy
• General Information about survey “microcensus”
• Variables by thematic classification and year
• List of variables by year
• List of generated variables by year
• Show details of a variable with statistics
• Variable-Time Matrix
• List of variables by thematic classification and year (selectable)

• Questionnaire Catalogue
49

Use-Cases in Missy

• In future: browse by survey and country!
50

Use-Cases in Missy

51

Requirements to Software Developers
• Complex data model to represent use-cases
• Focus lies on reusability
• Flexible Web Application Framework and modern architecture
• Service-oriented
• Semantic Web technologies
• Creation of an abstract framework and architecture
• should be well designed to be able to be extended and reusable

57

What comes next?
• We have seen Missy as an use-case with several fields which
holds values

• I would like to use DDI
• What are the main structures of DDI-L, Thomas?

58

What comes next?

• DDI Instance
Presentation
• Study Unit


59

What comes next?

• DDI Instance
Presentation
• Study Unit


63

ResourcePackage

Allows packaging of any maintainable item as a resource item
Structure to publish non-study-specific materials for reuse
67

ResourcePackage

Any module (except: StudyUnit, Group, LocalHoldingPackage)
68

ResourcePackage

Any Scheme
Internal or external references to schemes and elements within schemes
69

Group

Groups allow inheritance
71

LocalHoldingPackage

75

LocalHoldingPackage

• Add local content to a deposited study unit or group
• without changing the version of the study unit, group
76

What comes next?

• DDI Instance
Presentation
• Study Unit


77

TemporalCoverage

83

TemporalCoverage

88

FundingInformation

89

FundingInformation

90

How the presentation is continued
• Ok, I have shown you
• the main structures as well as
• the main modules of DDI-Lifecycle

• If you want to use DDI you have to have a model or idea of
your software architecture
• How would your software architecture look like?

95

What comes next?

• DDI Instance
Presentation
• Study Unit


96

What makes Missy interesting
• Cross-linking between different multilingual studies
• enabled by the common data model
• provides different use-cases
• Missy leverages DDI as its back-end data model and exchange
format
• Modern web project architecture, e.g. Multitier, MVC, etc.

• Is designed to be published as open-source project with an
API to persist DDI data

97

Software Architecture
• Standard technologies to develop software
• a multitier architecture
• Model-View-Controller (MVC-Pattern)
• Project management software, e.g. Maven

• Multitier architecture separates the project into logical parts
• presentation
• business logic or application processing
• data management
• persistence
• …
98

Model-View-Controller
• separation of
• representation of information
• interactions with the data
• components

• the key is to have logically separated parts, where people
might work collaboratively

100

MVC – Interactions

View
controls

Controller accesses

manipulates

Model

101

Model
• Represents the business objects, i.e. real life concepts, and
their relations to each other
• Sometimes also includes some business logic

• Independent of the presentation and the controls

102

View
• Responsible for the presentation of information to the user
• Provides actions, so that the user can interact with the
application

• "Knows" the model and can access it

• Usually does not display the whole model, but rather special
"views" or use-cases of it

103

Controller
• Controls the presentations, i.e. views
• Receives actions from users
• interprets/evaluates them
• acts accordingly

• Manipulates the data / model
• Usually each view has its own controller

• Be warned of code reuse of controllers
• controllers are most often the throwaway components
104

View Technologies

View
controls
…

Controller accesses

manipulates

Model

105

Data Model

View
controls
…

Controller accesses

manipulates

Model

106

Missy Technologies

View
controls

Controller accesses

manipulates

Model

107

What we have discussed so far
• Brief introduction to Missy and the use-cases
• With which technologies a web application framework might
be build up today
• How a modern software architecture looks like

108

• What the DDI-L core modules are and how they are organized

• Introduction the disco model and
• How to extend it

109

What comes next?

• DDI Instance
Presentation
• Study Unit


110

ConceptualComponent

ConceptualComponent

111

ConceptualComponent

112

ConceptualComponent

113

What comes next?

• DDI Instance
Presentation
• Study Unit


120

CodeRepresentation

138

CodeRepresentation

139

NumericRepresentation

140

NumericRepresentation

141

TextRepresentation

142

TextRepresentation

143

What comes next?

• DDI Instance
Presentation
• Study Unit


146

CollectionEvent

152

CollectionEvent

153

Response Domain

158

Response Domain

159

ControlConstruct

162

ControlConstructScheme

163

ControlConstruct

164

Sequence (bahavioral)

166

Sequence (XML)

<Sequence>
<ControlConstructReference> [Q1] </ControlConstructReference>
</Sequence>

167

Sequence (structural)

168

StatementItem (behavioral)

170

StatementItem (XML)
<StatementItem>
<DisplayText>
statement
</DisplayText>
</StatementItem>

171

StatementItem (structural)

172

QuestionConstruct

173

QuestionConstruct (behavioral)

174

QuestionConstruct (XML)
<QuestionConstruct>
<QuestionReference> [Q1] </QuestionReference>
</QuestionConstruct>

175

QuestionConstruct (structural)

176

IfThenElse (behavioral)

178

IfThenElse (XML)
<IfThenElse>
<IfCondition>
<Code programmingLanguage="…"> yes </Code>
<SourceQuestionReference> [Q1] </SourceQuestionReference>
</IfCondition>
<ThenConstructReference> [Q2] </ThenConstructReference>
<ElseIf>
<IfCondition>
<Code programmingLanguage="…"> no </Code>
</IfCondition>
<ThenConstructReference> [Q3] </d:ThenConstructReference>
</ElseIf>
</IfThenElse>
179

IfThenElse (structural)

180

Loop (behavioral)

182

Loop (XML)
<Loop>
<InitialValue>
<Code programmingLanguage=“Java"> i = 0 </Code>
</InitialValue>
<LoopWhile>
<Code programmingLanguage=“Java"> i < 10 </Code>
</LoopWhile>
<StepValue>
<Code programmingLanguage=“Java"> i = i + 1 </Code>
</StepValue>
</Loop>

183

Loop (structural)

184

RepeatWhile (behavioral)

186

RepeatWhile (XML)
<RepeatWhile>
<WhileCondition>
<Code programmingLanguage=“Neutral"> no </Code>
<SourceQuestionReference> [Q1] </ SourceQuestionReference >
</WhileCondition>
<WhileConstructReference> [Q1] </WhileConstructReference>
</RepeatWhile>

187

RepeatWhile (structural)

188

RepeatUntil (behavioral)

190

RepeatUntil (XML)
<RepeatUntil>
<UntilCondition>
<Code programmingLanguage="Neutral"> yes </Code>
</UntilCondition>
<UntilConstructReference> [Q1] </UntilConstructReference>
</d:RepeatUntil>

191

RepeatUntil (structural)

192

What comes next?

• DDI Instance
Presentation
• Study Unit


193

How to extend the disco model?

195

What we have seen so far
• An introduction to the core DDI modules

• A short introduction to the disco-model, which covers the
most important parts of DDI-Lifecycle and DDI-Codebook

196

• Now that we know the data model…

• How can it be used for the Missy use-case, Matthäus?

• Can we create a mapping of the fields in the disco-model and
the fields in Missy?

197

What comes next?

• DDI Instance
Presentation
• Study Unit


198

Use-case “variable details”

199

Sub-Sample:
• is-a Sub-Universe with the
UniverseType „subSample“
Filter Statement
• is-a Sub-Universe with the
UniverseType „filterStatement“
205

• disco-model
• designed for the discovery use-case
• provides object types, properties and data type properties designed
for discovery use-case

• One Missy use-case and all the data type and object
properties that are covered

• The idea of how theoretically the disco can be extended

214

• But, what if a project has specific requirements to the model,
that are not covered by the disco model?

• How are we going to implement that?

215

What comes next?

• DDI Instance
Presentation
• Study Unit


216

disco-model API
• The disco-model implementation is designed to be extended
• For one project’s individual requirements

• Speaking in MVC-pattern language: the model is separated
and independent from the views and the controllers

220

disco-model API
• abstract classes
• define the basic elements, i.e. data type and object properties, that
are provided and covered by the disco-model
• may be extended, but cannot be instantiated
• e.g. AbstractVariable
• model classes
• represent the actual entity types of disco-model
• may be instantiated and reused
• e.g. Variable

221

disco-model API
• abstract classes
• define the basic elements, i.e. data type and object properties, that
are provided and covered by the disco-model
• may be extended, but cannot be instantiated
• e.g. AbstractVariable
• model classes
• represent the actual entity types of disco-model
• may be instantiated and reused
• e.g. Variable

• Implementation of the Template Pattern
222

Project specific extensions

225

Project Specific Extensions
• Projects may just simply use
• and extend the defined classes
• Introduce new fields and methods, such as getter/setter
• Change the type of existing fields when they overwrite the return type
(in Java this is only possible in case of covariant returns)

226

How is the presentation continued
• Now that I have the data model… and I can extend it
• I just need to store my data somewhere…

• Are there any persistence mechanisms that are offered by the
DDI-Lifecycle model?

227

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

228

PhysicalDataProduct

229

PhysicalDataProduct

230

PhysicalDataProduct

231

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

232

PhysicalInstance

233

PhysicalInstance

234

PhysicalInstance

235

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

236

Archive

Contains persistent lifecycle events
Contains archive-specific information (archive reference, access, funding)
239

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

240

DDIProfile

Describes which DDI-L elements you use in your institution’s DDI flavor
243

• How can the disco-model be persisted?

and

• How can it be done in a project with individual data model
elements?
• How do we do this in the Missy application?

244

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

245

Data Storage Access

246

Persistence Layer – Motivation
• In a well-defined software architecture the application itself
does not need to know how the data is stored
• It just needs to know the API
• Methods that are provided to access and store objects
• May be abstracted away from the actual implementation

• An actual implementation or strategy can just be a matter of
configuration

249

Persistence Layer – Strategy Pattern
• Implementation of the Strategy Pattern
• simple interface
• actual implementations are the different strategies

• A strategy is an implementation of the actual type of
persistence or physical storage, respectively
• e.g. DDI-L-XML, DDI-RDF, XML-DB, Relational-DB, etc.

250

Modules
• disco-persistence-api
• Defines persistence functionality for model components regardless of the
actual type of physical persistence
• disco-persistence-relational
• Implements the persistence functionality defined in disco-persistence-api
with respect to the usage of relational DBs
• disco-persistence-xml
with respect to the usage of DDI-XML
• disco-persistence-rdf
with respect to the usage of the disco-specification

252

Data Access Object
• A DAO is an object that is responsible for providing methods
for reading and storing of objects in our model
• A DAO is again implemented against an interface

• For each business object in the data model there exists a DAO

• Persistence strategy interface defines methods for obtaining
specific DAOs (data access objects)

253

disco-persistence-api
• Provides an API and implementations for the disco-model

• PersistenceService
• Available in the disco-persistence-api
• Encapsulates the actual strategy

• Strategy can be instantiated and injected into the PersistenceService

255

Extendable disco-persistence-api
• But, projects have own requirements to the data model
• => new business objects are included
• => DAOs need to be written for reading and storing these objects
• => these DAOs also have to implement the appropriate type of
persistence used by that project

• => this API has to be extendable

256

Extendable disco-persistence-api
• But, projects have own requirements to the data model
• => new business objects are included
• => DAOs need to be written for reading and storing these objects
• => these DAOs also have to implement the appropriate type of
persistence used by that project

• => this API has to be extendable

• Luckily it is, because of the interface structure!
• Projects only need to implement the defined methods
257

Missy Specific Modules
• missy-persistence-api
• Defines additional persistence functionality for model components
regardless of the actual type of physical persistence
• missy-persistence-relational
• Implements additional persistence functionality for model
components with respect to the usage of relational DBs
• missy-persistence-xml
components with respect to the usage of DDI-XML
• missy-persistence-rdf
components with respect to the usage of the disco-specification
260

• Implementation of a persistence API for the disco model
• How the persistence API can be extended in individual
projects

261

• How does, on the persistence level, the serialization of data
stored in DDI really look like?

262

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

263

Questionnaire - Question

264

Questionnaire – Question (DDI 3.1 XML)

265

DataCollection

Instrument
IntrumentName
ControlConstructReference

ControlCostructScheme
QuestionConstruct
QuestionReference

QuestionScheme
QuestionItem
QuestionItemName
QuestionText
266

Questionnaire – Question (DDI-RDF)

267

Variable - SummaryStatistics

268

Variable – SummaryStatistics (DDI 3.1 XML)
PhysicalInstance
Statistics
VariableStatistics
VariableReference
SummaryStatistic
SummaryStatisticType
'Minimum'
Value
10

Variable
VariableName
'v1'
Label
'Age' 269

Variable – SummaryStatistics (DDI-RDF)

270

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

271

Relational-DB Example

272


273


274

Inheritance of Properties
From Identifiable

275

Inheritance of Properties
From abstract class

276

Declaration of own Properties
Own properties

277

What comes next?
• What are you going to do with the MISSY project?

• Is there ongoing work?

278

Outline

• Archive
• Examples

• DDI-XML
• Relational-DB
Data Storage

279

Outlook
• YES! There is ongoing work!

• Use the API in other projects, e.g. GESIS internal projects like
StarDat
• Publish the persistence API as open-source on github

• Two presentations at IASSIST 2013
• One poster at IASSIST 2013

280

Poster
• Shows detailed information regarding Missy
• Multitier Architecture, MVC
• Employed Technologies
• Specific Maven Modules
• Project Home
• Excerpts from disco-model and the mapping of the fields

281

Thank you for your attention…
• Feel free to download the sources from GitHub:
https://github.com/missy-project

• Give us feedback!
• Feel free to criticize!

Thomas Bosch Matthäus Zloch
GESIS - Leibniz Institute for the Social Sciences GESIS - Leibniz Institute for the Social Sciences
thomas.bosch@gesis.org matthaeus.zloch@gesis.org
boschthomas.blogspot.com

283

2012.12 - EDDI 2012 - Workshop

More Related Content

Viewers also liked (10)

Similar to 2012.12 - EDDI 2012 - Workshop (20)

More from Dr.-Ing. Thomas Hartmann (20)

2012.12 - EDDI 2012 - Workshop