Andy Turner's MoSeS Meeting 2006-05-23 Page

Andy Turner's MoSeS Home Page

People

• Mark
• Phil
• Martin
• Andy

Ice Breaker

• Phil made coffee, thanks Phil!

Documentation

• Document Mark circulated
• Document Andy circulated

Agenda

• Go through document Mark circulated
• Discuss work that Mark has been doing with some Genetic Algorithms

Notes

• Meeting based on document Mark circulated
• Andy made notes while Mark ammended the document based on feedback from the group
  • Martin, Mark and Justin met yesterday to think about upcoming meeting with NHS
  • Belinda needs a baseline to be testing her parts of the model
  • Ethnicity is a very important factor that should be included in the model
  • What are we going to do in the next version of the model
    • Is going to MSOAs going to help with the SCAM problem?
      • Phil thinks yes
      • I think we don't know, but suspect that the aggregation is only beneficial as a reporting mechanism
    • We all agree that IPF is useful for adding variables from other data source e.g. BHPS
    • Jin has found lots of inconsistencies in the HSAR
      • these have been reported
    • I wonder about imputed variables in the HSAR
    • Discussed data cleaning
      • Think this is too much of a diversion at this stage...
      • May return to it later...
  • Plan to automate results production (tables, index of dissimilarity, mapping)
  • Phil wants to see a flow diagram based on the document he developed during my Staff Review

• Modelling Decisions

  • Apply age control totals (constraints) to HRPs only instead of total population
  • Use Household SAR exclusively for household populations
  • Continue to use ISAR for communal establishment populations
  • Constrain and optimise based on MSOA totals rather than OA totals
  • Aggregate to MSOAs for reporting purposes

Actions

• Mark to email Belinda and ask about what she needs
  • This is already clear based on 2006-05-24 IRC between Belinda and Andy
  • Andy to speek to Belinda and confirm... done... nothing has changed... great!
• Andy to advise Mark what to write in para 6 Section 3.2 of this document
• Andy to develop flow diagrams:
  • UML of MoSeS code
  • UML flow diagram of MoSeS code
  • Descriptive flow diagram outlining the population initialisation process
• Andy to develop a new version of the population initialisation process based on the Modelling Decisions agreed in this meeting...