Temporal GIS and Statistical Modelling of Personal Lifelines

1
Temporal GIS and Statistical Modelling of
Personal Lifelines

• Marius Thériault, Christophe Claramunt,
  Anne-Marie Séguin and Paul Villeneuve

July 2002 Spatial Data Handling Ottawa
Funded by the Canadian SSHRC, the Canadian NSERC,
the Quebec Provinces FCAR and GEOIDEs project
SOC8
2
Introduction

• Urban modelling must consider decision-making
  behaviour of urban actors using disaggregate data
• Activity location, home choice, commuting and
  travel decision
• Household and professional profiles of persons

• Probabilistic discrete-choice theory is becoming
  the central tenet of urban modelling research
• Implemented using logistic and Cox regression
  techniques
• Aimed at modelling individuals and households
  behaviour

• Needing dynamic spatial tools for analysing
  complex urban systems where
• Uncertainties exist in the system (aggregation is
  not straightforward)
• Decision rules for individuals and households can
  be intricate
• System processes are path and location
  dependent- future system state depends partly on
  past and current states (thus needing event
  history analysis)

3
Purpose and Objectives

• Purpose
• Detect the unintentional consequences, at the
  macro scale (E.g. urban spread), of intentional
  actions and strategies occurring at the
  micro-scale (statistical aggregation)

• Objective develop a logical database model to
  handle personal biographies and to restructure
  individual lifelines data in a format suitable
  for statistical analysis
• Needing to build a new spatio-temporal dataset
  (flat file) for any question at hand (data
  restructuring for statistical analysis)
• GIS needed to study influence of neighbourhood
  on individual decisions and to summarise their
  combined effect on the evolution of the overall
  urban system

4
Why Studying Individual Biographies?

• Focus of this work
• Household, residential and professional history
  of citizens

• Life course of most individuals
• Is built around three interlocking series of
  events- a household history an occupational
  career a residential trajectory
• During the last decades, these trajectories
  generated patterns of events of increasing
  complexity- more divorces- extension of
  contractual short-term employment- increasing
  geographical mobility, etc.
• Within cities, these individual trajectories
  intersect and combine, yielding demographic and
  residential patterns driving city evolution and
  transportation demand

• Understanding processes by which personal
  biographies aggregate and evolve cannot be
  derived from censuses
• They give only the barest spatio-temporal
  snapshot reports on complex situations and they
  do not relate facts

5
Example of an Individuals Biography
6
Changes in Personal Life

• An individuals history is altered
• When an event occurs modifying at least one
  important aspect of his personal status (marital,
  family, job, home, education, income, etc.)
• Such an event may alter simultaneously status on
  more than one trajectory - or have effect on
  several individuals in the family
• Some events (E.g. new born baby) can be
  anticipated and may potentially lead to prior
  adjustment (actions linked to expectation)
• Effects can also be delayed (after the enabling
  event occurs)

• Life trajectories show interlocked evolution
• Behaviour based on personal values, beliefs and
  strategy
• They associate episodes (time periods with stable
  attributes) which intersect to depict global life
  status of the person
• Hypothesis their ordering builds logical
  sequences (evolution patterns) related to life
  cycles (E.g. young couples, retired persons,
  etc.)
• Studying these patterns is more relevant to urban
  studies than knowing the exact timing of events
  for each individual

7
Using Retrospective Surveys

• Retrospective surveys
• Provide detailed information about changes
  occurring during the life of the respondents
• However, this spatio-temporal data must be
  properly structured and carefully analysed to
  reveal spatio-temporal structures and patterns

• Advantages
• Phenomena are measured for individuals
  (micro-level)
• The follow-up cover long periods of time (E.g.
  since birth, marriage or departure from the
  parents home)
• Information can be structured using lifelines and
  personal trajectories

• Specific issues
• Data reliability and questionnaire structure -
  respondents have to remember places and events
  that were happening many years earlier
• However, sequence of events are more reliable
  than dates
• Spatial and temporal data may be fuzzy
• Need appropriate data modelling to handle
  historical sequences and to allow comprehensive
  time-based statistical analysis

8
The 1996 Retrospective Survey for Quebec City

• In Quebec City, a retrospective survey
  collecting, in one interview, information about
  all changes occurred over a long period of time,
  since the departure of the parental home
• A spatially stratified sample of two cohorts of
  professional workers
• Sample of 418 respondents stratified by
  municipality, gender and age cohort (36-40 and
  46-50).
• Interviews realized at the respondents home,
  mean duration 1.5 hour
• Three trajectories
• Residential trajectory every home occupied
  (three months or more) since the departure of
  parents home, with their location (civic
  address) and other characteristics (tenure,
  price, choice criteria, reasons to leave, etc.)
• Household trajectory each change in the
  composition of the respondents household
  (arrival or departure of a spouse, birth, death,
  arrival of a child from an other household,
  relatives, roommates, cotenants, etc.)
• Professional trajectory each change in
  employer, each work place, with their
  characteristics (including secondary jobs,
  education and unemployment episodes)
• Collecting dates of every change (starting- and
  ending-time of each episode)

9
Spatio-Temporal Modelling of Biographies

• Relate to the integration of time in GIS
• Triad framework proposed by Peuquet (space - time
  - theme)
• Integrate the notion of event-process and
  jointly-related entities described in Claramunt
  et al.

• Main task
• Design a relational database schema of
  individuals trajectories and providing query
  mechanisms needed to restructure spatio-temporal
  data in a format suitable for statistical
  analysis (using GIS and DBMS)

• Main characteristics
• Entity-based implementation within RDBMS and GIS
  to describe individuals, events, processes,
  households, jobs, diploma, etc.
• Building multi-dimensional sequences of events
  combining lifelines and trajectories
• Providing flat files needed for statistical
  analysis of ad hoc queries

10
Modelling Life Trajectories

• Specific conceptual modelling issue
• How can we express the temporal structure of
  biography as an ordered sequence of intertwined
  statuses and events, using database modelling
  concepts, while retaining its behavioural meaning?

• Personal biographies
• Are a complex mix of real world phenomena (E.g.
  persons, dwellings, etc.) generating abstract
  temporal features (E.g. episodes, events)
• Episodes are ordered along lifelines to form
  sequences of independent or joint evolution
  (linked trajectories or related individuals)
• Trajectories hold sets of relationships
• Aggregation (household made of persons),
  combination (mix of jobs held simultaneously), or
  collaboration (renting or buying a dwelling is
  using another type of entity and starts a new
  residential episode)

11
Database Modelling of Trajectories

• Modelling concepts
• Trajectories are combining events and episodes
  describing a multi-dimensional aspect of personal
  life
• Each trajectory (E.g. household) groups a set of
  related lifelines (E.g. marital status, family
  composition)
• Each lifeline describes a specific dimension of a
  trajectory, ordering episodes (periods of time)
  during which a given status was stable (E.g.
  single or married).
• When an event occurs, there is some change in
  status, leading to a new episode (E.g. birth of a
  child in an household changes its composition)
• Events and episodes form sequences ordered along
  lifelines (directional from past to future)

12
Quebec City - Trajectories and Lifelines

• Our survey questionnaire leads us to define nine
  lifelines related to the three trajectories
• The first lifeline is used to depict the
  Respondent life (from birth up to the interview)
• Residential trajectory
• Is simple its a formed by one lifeline Home
  tenure episodes and home related events (rent,
  buy, sell, etc.)
• Household trajectory
• Is more complex it relates 3 types of lifelines
  Marital status (including identification of
  successive spouses, union, separation) Children
  (ordering events and episodes related to children
  birth, adoption, departure after a divorce)
  Household lifeline makes the synthesis of any
  family change
• Professional trajectory
• Is complex it relates 4 types of lifelines
  Educational (including degrees), Occupational
  (mix of independent or simultaneous jobs) and
  Work place histories They combine to form the
  Professional lifeline

13
Spatio-temporal Database Modelling

• The core of the spatio-temporal model is formed
  by an EPISODE table combining events and episodes
  ordered by the EpisodeSequences table.
• The ontology of lifelines and trajectories is
  stored directly in the database, and each event
  and episode is typified
• Episodes are related to the respondent, but also
  to any other acting individual in the household
  (E.g. linking the respondent to his/her spouse)
• Every episode and event is related to space
  through the SPATIAL table providing locations
  managed by Spatialware functionalities (E.g. ODBC
  link with Access)

14
Modelling of Residential Trajectory

• The Residential trajectory is made of only one
  lifeline describing the attributes and occupation
  modes of successive homes inhabited by the
  respondent during his/her adult life
• Each home is located in space using street
  addresses and location is managed, for each
  episode, by the SPATIAL table
• Integrity constrains are enforced
• GIS operations are realized within MapInfo using
  ODBC technology with Spatialware features
  generating points in Map Views

• The Home lifeline handle home tenure related
  events (Rent, Buy, Inhabit) and episodes (Tenant,
  Owner, Cotenant, etc.)

15
Modelling of Household Trajectory
16
Modelling of Professional Trajectory
17
Querying Using Temporal Sequence Views

• These trajectories and lifelines are related into
  a unified database structure describing their
  successive temporal, spatial and thematic
  attributes
• The relational model allows for building
  relationships across lifelines, events and
  trajectories using state transition views similar
  to the example shown here
• These integrated views are later used to ease
  formulation of spatio-temporal query building ad
  hoc event history datasets needed for statistical
  analysis

Episode Event - Episode
Event Episode - Event
18
Linking to Event History Regression Analysis

• Most of the phenomena discussed in this research
  may be thought as events
• These events and their possible relationships are
  recorded using RDBMS
• We want to submit to statistical analysis these
  data and expressions based on them in order to
  build event history models
• Ordinary multiple regression is ill-suited to the
  analysis of biographies, because of two
  peculiarities censoring and time-varying
  explanatory variables

• Censoring refers to the fact that the value of a
  variable may be unknown at the time of survey,
  generally because the event did not occur (E.g.
  duration of marriage for a person who never
  divorce)

• Considering time varying explanatory factors
• To study the effect of the family composition on
  residential location choice, one needs to
  consider time-varying information
• A bio-statistical method called event history
  regression analysis can handle such a problem (it
  combines survival tables and logistic regression)
• Our approach enable data restructuring that
  fulfil requirements for this kind of statistical
  analysis

19
Event History Analysis

• Survival tables are using conditional
  probabilities to estimate the mean proportion of
  people experiencing some change in their life
  after a significant event occurs (E.g. proportion
  of tenants buying a home after the arrival of the
  second child), computing the time delay after a
  specified enabling event (E.g. time to divorce
  after marriage)
• However, these probabilities are not exactly the
  same for everyone because specific conditions may
  influence propensity to change
• Finding those specific factors that condition
  individual propensity to do something requires a
  combination of survival tables and logistic
  regression in order to estimate the marginal
  effect of other personal attributes on the
  probability that an event occurs
• The purpose of Event History Analysis (also
  called Cox Regression) is to model specific
  variations of the probability of state transition
  through time for individuals considering
  independent (even time-varying) variables
  describing their personal situation on other
  lifelines (E.g. What is the marginal effect of a
  6-month unemployment period occurred less than
  five years ago, on the propensity to buy a home
  after the second child is born? Is their a
  significant effect? Is this effect stable over
  time and space?)

20
Proportions of Tenants and Home-Owners Related to
Time Elapsed Since Departure From Parents Home
(survival rates)
Cohorts 1 respondents in their thirties 2
respondents in their forties
Proportion of tenants
Proportion of home-owners
Cohort
Time elapsed since departure from parents home
(years)
Cohort
Time elapsed since departure from parents home
(years)
21
Proportions of Tenants and Home-Owners Related to
Respondents Age (survival rates)
Cohorts 1 respondents in their thirties 2
respondents in their forties
Proportion of tenants
Proportion of home-owners
Cohort
Respondents Age (years)
Cohort
Respondents Age (years)
22
Spatio-temporal Analysis of Retrospective DataAn
Example

• Study individual behaviour of persons making
  home-location choices under various conditions
  linked to their own personal history, considering
  the three trajectories
• The retrospective survey and its implementation
  within a temporal GIS providing event-ordering
  functionalities will further our understanding of
  their strategies for moving through the city
  considering their own history, the impact of
  growing family, of changes in work place, their
  educational status, income, home price, stability
  in employment, etc.
• The next slide shows a preliminary event-history
  model of the propensity of tenants to buy a house
  after the birth of their first child
• It was realised with a very preliminary version
  of the spatio-temporal database, but can help
  understand the advantages of such an approach

23
Event History Analysis Results

• What are the factors influencing the decision of
  tenants to buy a house (with some delay) after
  their first child is born?
• Is there significant differences among persons?
  Yes (Chi Square)
• Are their behaviour stable over time? No
  (Significant variations of propensity over time)
• What are the factors influencing the propensity
  of changing status from tenant to home-owner? 1-
  Stability in employment, 2- Decade during which
  the child was born (time varying behaviour), 3-
  Willingness to move towards remote locations
  (related to house prices)

24
Event History Statistical Model
Example of Application in Excel (Activate the
Spreadsheet to set parameters)
25
Event History Analysis Results
Stability in employment increases propensity to
buy a home
Rate of access to property ownership
significantly increases through time - from the
sixties to the eighties
26
Discussion and Conclusion

• The proposed database modelling approach is
  classical
• It uses entity-relationship principles, combined
  with geo-relational technology
• With the exception of minor enhancements to
  existing methods, the purpose was not to
  contribute to the advance of STDB modelling

• Contribution
• To the best of our knowledge, this type of
  application for the spatial monitoring of changes
  in population behaviour is original
• Keeping track of dynamics within GIS database is
  a very important requirement for urban and
  transportation planning
• Towards a generic spatio-temporal life
  trajectories ST schema

• Multidisciplinary approach
• Combining TGIS and event history analysis provide
  methodology well suited for behavioural modelling
  applications
• We are now developing application-independent
  dialogs for querying this spatio-temporal
  structure and build projections yielding flat
  files needed for statistical analysis using SPSS,
  SAS or SPlus