2016-04-06

[andkov]

Agenda:

1. debug multistate repo with github client
2. discuss new outcome space map
3. Draft the table of events for the Victoria workshop

[andkov]

During meeting

• what if there are more one specific test to a general operationalization

program_name sucommand/argument

git status

pulls from the cloud repository and tries to merge things

git pull

pull is a combined arguments: fetch and merge

difference b/w two documents, varies what they are

git diff

Victoria workshop

• 2.5 days
• cognitive-physical focus
• be conservative. decide what is most important then cut in half. call the other half "optional".
• the first one was to inefficient.
• we want something more scalable,
  • less work for them
  • less risk (of manual errors) for the modellers
  • quicker publishable product is done within day.
• insist to communicate that the main objective is SCALABILITY

• Make wide data happen before Day 1.

  DAY 1 ,

• (1) Adjust data set to be used by PCS
• (2) Complete PCS; / lunch
• (3) Begin estimating the models

• (4) Estimate the models

  DAY 2

• (1) Continue modelling
• (2) Start interpreting forest plots lunch
• (3) begin interpretation

• (2) continue interpretation

  Day 3

• (1) discuss finding

• (2)write report

  ELSA

  Homework for Maleeha Use ELSA data

• Complete PCS with respect to variables existing in ELSA.

[andkov]

Specification of the WIDE data to be fed into PCS

Static variables

• a.k.a time-invariant variables
• values that don't change over the life of the subject
• e.g. subject_id, dob, gender

Dynamic variables

• time varying variables
• if there are k time-variant measures, there will be at most k+1 variables per time point (additional measure capturing subject's age at the time of the measure.

Specific requirements

• must be a csv with column names
• missing values are indicated by a formal NA, not a numeric code (e.g. 9999)