This zip file contains two Powerpoint presentations by Martin Spielauer (World Bank consultant), dated February 2017 (presentation at University of Southampton) and May 2017 (presentation at the Mauritanian Center for Policy Analysis). The presentations cover the rationale and methods of dynamic micro-simulation, and describe a portable population projection application.
Dynamic micro-simulation for population projections
This project aimed to demystify the use of dynamic micro-simulation for population projections and to demonstrate its feasibility and relevance in the context of developing countries. Advanced and freely available programming tools and improvements in the availability and quality of micro-data have helped make dynamic micro-simulation feasible at a reasonable cost. At the core of this project is an application example for Mauritania, which introduces statistical methods used in micro-simulation and micro-simulation programming techniques, and can be adapted and replicated for other countries. The model has an intuitive graphical user interface and runs on a standard personal computer. Its code and statistical analysis files are available to model builders, who can use them as a textbook and toolbox for micro-simulation model development and implementation.
| Project status: | Closed (but follow-up activities are being implemented) |
| Related on-going or planned activities: | The project is being extended to a second country, with a new focus on projecting populations by education level and on geo-referencing the projected populations. |
| Sponsor(s): | DFID Trust Fund No TF011722 administered by the World Bank; Knowledge for Change Program (KCP III) - Grant No TF0A1095 administered by the World Bank; World Bank budget |
| Implemented by: | World Bank Development Data Group and Centre Mauritanien d'Analyse des Politiques (CMAP) |
| Type of output: | Reproducible and adaptable population projection model by dynamic micro-simulation, developed using MODGEN (freeware by Statistics Canada), and detailed guidelines |
Output
Martin Spielauer, Olivier Dupriez -
Powerpoint presentation, presented at the 6th World Congress of the International Microsimulation Association (Torino, Italy, 21-23 June 2017).
May 2017 (version v01)
This report describes in detail the rationale, approach of micro-simulation applied to population projection. It provides all information needed to modelers interested in replicating or adapting the model developed in collaboration with the Mauritanian Center for Policy Analysis (CMAP) and Office of National Statistics (ONS). The model is named DYNAMIS-POP (for DYNAmic MIcro-Simulation - POpulation Projection). This report is related to version 1 of the model (DYNAMIS-POP-MRT-2017_v01). The source code of the model, the parameters tables, the Stata scripts and Excel files produced to generate the parameters tables and scenarios, are also made publicly available. The model is developed using MODGEN (MODel GENerator, a freeware by Statistics Canada).
May 2017 (version v01)
This report describes in detail the rationale, approach of micro-simulation applied to population projection. It provides all information needed to modelers interested in replicating or adapting the model developed in collaboration with the Mauritanian Center for Policy Analysis (CMAP) and Office of National Statistics (ONS). The model is named DYNAMIS-POP (for DYNAmic MIcro-Simulation - POpulation Projection). This report is related to version 1 of the model (DYNAMIS-POP-MRT-2017_v01). The source code of the model, the parameters tables, the Stata scripts and Excel files produced to generate the parameters tables and scenarios, are also made publicly available. The model is developed using MODGEN (MODel GENerator, a freeware by Statistics Canada).
May 2017
This file contains all Modgen files (model code and scenario) needed to replicate the production version of the DYNAMIS-POP-MRT-2017_v01. Another file is also made available on this web page, which provides a step-by-step version of the development of the model (in 19 steps).
We also provide in this web page a USER version of the model (executable Modgen model), to be used by those interested in implementing the model with the parameters we provide, or with new scenarios/parameters.
May 2017
Executable version of the model DYNAMIS-POP-MRT-2017_v01. This collection of files allows replication of the model. A baseline population of 100,000 individuals, and all parameters and scenarios, are provided. Running the executable version of the model requires prior installation of the MODGEN Prerequisite software (freeware by Statistics Canada; link provided in this web page).
Martin Spielauer (World Bank consultant) - May 2017
This zip file contains all source code (Modgen language) used for the development of the DYNAMIS-POP model developed for Mauritania in 2016/2017. The model was developed in 19 steps, starting with the replication of a simple cohort-based population projection model, then progressively adding complexity showcasing the power of the micro-simulation approach.
A technical report is also available, which provides a very detailed description of the modules, their purpose, and their design.
May 2017
This file contains all parameters (Excel files) used to run the micro-simulation population projection model developed for Mauritania (DYNAMIS-POP-MRT-2017_v01).
The Excel files can be used to modify the assumptions made to run the model. They also provide charts and tables describing the parameters.
May 2017
This zip file contains a collection of Stata (version 14) do files used to analyze data from the Mauritania Population Census 2013 and Multiple Indicators Cluster Survey (MICS) 2015. The log files resulting from running the do files on the source datasets are also provided. This analysis of the census and survey data was conducted to genrate a starting population and to derive parameters (core and relative risk tables) for the population projection model. The methods applied in these Stata scripts include simple tabulations, and various regression models (including proportional hazard regression models).
NOTE: The source datasets for these scripts are the property of the Office of National Statistics (ONS) of Mauritania. The scripts that require access to the Population Census data were run by ONS staff. Other scripts apply to the MICS 2015 survey dataset, which is (or will be) made publicly available on-line by UNICEF (visit http://mics.unicef.org/surveys).
Statistics Canada -
The model DYNAMIS-POP-MRT was developed using Modgen 12 (Model Generator), a generic programming language (freeware) supporting the creation, maintenance and documentation of dynamic micro-simulation models. Modgen is developed and disseminated by Statistics Canada.
Two components need to be downloaded to use Modgen: Modgen itself (to develop or edit models), and Modgen Prerequisites (to run models as executable files).
Developing or adapting Modgen models require a C++ compiler; such as the free Microsoft Visual Studio Community Edition 2017.