Data Modelling

Dennis has been involved with data and financial modelling for many years.  His most recent work has been to provide some of the statistical support for United Way, especially the Poverty Solutions work.  Statistics Canada’s standard geographic based data do not match well with local neighbourhoods of interest.  As a result, considerable modelling of the Census data is needed to develop neighbourhood profiles. UW 2016 Census Snapshot  is an example of the type of work provided to United Way.

Because of the complexity of the data modelling, it was felt that the process should be documented.  The original concept was to document and explain the modelling approach, challenges and limitations of modelling the Census of Population data for United Way’s three “Action” neighbourhoods.  During this work, two other needs or opportunities were identified:

  • The DA boundaries established by Statistics Canada were not always appropriate, e.g. cutting through buildings, not following property lines and not reflecting local population patterns. There is a need to review the worst of these and suggest changes to Statistics Canada.  The window of opportunity for the 2021 Census may be passed, but any work on this can be incorporated into a future census if not this one.
  • HRM civic address details include the number of residential units. When comparing these with the Statistics Canada dwelling counts, there were many discrepancies, often large.  The reasons for the differences need to be explored and understood.  There is potential to use this analysis to determine any under-coverage issues that affect population counts.  This is important because many federal and some provincial funding programs are based on population counts derived from the Census.

Fairview Strange 2016 DA

A first draft of the paper, under development, provides examples and some preliminary research on both of these issues.  Additional work needs to be done to complete the paper over the next few months.

The following are examples taken from some of the appendices in the report.  The picture to the right is explained in Appendix B pt 2 below.

  1. Appendix B – Strange DAs and DBs
  2. Appendix B pt 2 – Strange DAs and DBs
  3. Appendix D – HRM and Stat Can Data Mismatch

This page last updated: April 15, 2018