Health

Research articles · Aug 31, 2017

simSALUD: Design and implementation of an open-source wizard based spatial microsimulation framework

Melanie N Tomintz, Bernhard Kosar, Victor M García-Barrios

GeoHealth Laboratory, University of Canterbury, New Zealand; Carinthia University of Applied Sciences, Austria; Galileo University, Guatemala

Volume 10(2) Summer 2017

Cite as: M. Tomintz, B. Kosar, V. García-Barrios; 2017; simSALUD: Design and implementation of an open-source wizard based spatial microsimulation framework; International Journal of Microsimulation; 10(2); 118–143. doi: 10.34196/ijm.00159

Abstract

Spatial microsimulation models have been in use for around 40 years for various applications. Responding to users’ needs and project purpose, many have emerged using different methods and algorithms. Despite this, the literature suggests a gap in applications combining operability, flexibility, availability or generally an implementation of an online web mapping application in the area of spatial microsimulation. In this paper we introduce the development of a new open-source web-based spatial microsimulation framework, called simSALUD, as a solution approach for static simulation modelling where no programming skills are required. The emphasis lies on the design and implementation of the wizard and web-based spatial microsimulation application, including integrated validation and mapping elements. The current version supports two deterministic spatial microsimulation algorithms (combinatorial optimisation and iterative proportional fitting). The application is designed for experts as well as for non-experts to simulate their own spatial microdata. In addition, the application integrates statistical tools to validate and verify the robustness of the models ad hoc. The outputs of the simulation and validation results can either be visualized in the form of a map within the application or exported as comma separated files for further analyses. simSALUD comes with a demonstration dataset, an extensive documentation and video tutorial to disburden the usage of the software. The architecture of simSALUD was designed with the possibility to expand the tool towards a spatial decision support system, allowing measurement of possible future policy interventions on population groups at small geographies.

This article contains 5 figures and 2 tables.

Figures

Figure 1. The core components of simSALUD dedicated to the target audience (experts, non-experts, and academics) (see also Tomintz & García-Barrios, 2017).
Figure 2. Workflow of the spatial microsimulation model using the simSALUD application.
Figure 3. Struts 2 Model-View-Controller is realized by three application modules (Brown, Davis, & Stanlick, 2009).
Figure 4. System architecture of the simSALUD application.
Figure 5. Server architecture with an interaction process to visualize simulation data.

Tables

Table A.1. Functional requirements for the simSALUD framework.
Functional-Requirement Functional-Requirement Description
Data upload Users can upload their prepared .csv files through the application. The upload will be divided into two data parts: small area spatial data files and individual based files, e.g. survey data.
Choice of simulation algorithms Users can choose one of the implemented algorithm before starting the simulation process.
Data match of input files Spatial simulation requires a match between census and survey data. Therefore, an easy to use data match needs to be implemented, e.g. using drag and drop.
Contextual help User receives context-sensitive help that describes one current state, situation, or feature by using hover/tooltip.
Simulation progress During the simulation process, users should be able to follow the progress, e.g. progress bar, and follow the steps of the simulation process, e.g. information window.
Safe output data files The simulation and validation results can be saved as a .csv files by downloading it to the users local drive.
Map visualization A map window visualizes the simulated and validated data according to the chosen constraint straight after the simulation/validation execution.
Upload data for map visualisation Users are able to upload previous simulated and validated data for mapping them.
Save/Print customized map as pdf A pdf opens and the map can be printed or saved with the customized title, description, background map, etc.
Responsive for desktop, tablet The application should be responsive for desktop computers/laptop and tablets. Mobile is low priority as its use for this purpose is questionable.
Table A.2. Non-functional requirements for the simSALUD framework.
Non-Functional-Requirement Non-Functional-Requirement Description
Help / tutorial A help/tutorial is required to guide users through simSALUD and include all relevant information.
GUI (Geographical User Interface) evaluation A GUI evaluation is necessary using different target groups and to iterate this process after GUI improvements.
Expandability The architecture of the application needs to be designed for future integration of functionalities, such as more algorithms or geospatial processing tools.
Accessibility The application should be freely available and accessible via a web-based application.
Easy to use wizard Users are guided through the application by using a wizard design approach.
Well-designed A good design so that users find what they want and stay on the page to take actions.
Stability Avoiding system failures and provide a robust model application.