orange-system-dynamics 1.0.0.dev1

Add-on containing widgets for system dynamics operations

Orange3 System Dynamics Add-on

This is an add-on for Orange3 allowing to perform simulation using System Dynamic Modeling (SDM) mechanism.

Orange Data Mining

ODM (Orange Data Mining) is an open source machine learning and data visualization tool. It allows to build data analysis workflows visually, with a large, diverse toolbox.

We develop a list of widgets to be used in ODM, each widget has a functionality to be mainly applied for System Dynamic Modeling.

Installation

Orange3 and the SDM Addon must be installed in the same Python environment.

Orange 3

pip install pyQt5 PyQtWebEngine
pip install orange3

System Dynamics Addon

pip install orange-system-dynamics

Usage

Orange3 can be run using the following command:

orange-canvas

or

python -m orange.canvas

New widgets should appear in the toolbox bar under the System Dynamics section.

Widgets Description

Load Model Widget

Load Model widget is used to load simulation models in .mdl or .xmile format, it outputs the model for Simulation.

Simulation Widget

Once the model is loaded, the Simulation Widget is updated to show model's variables grouped by:

• Time Controls: contains the initial time, the final time and defined time step
• Stocks: contains all model’s stocks, can be used to change any stock’s initial value
• Auxiliary Values: contains all the auxiliary values, can also be changed

Once the Run Simulation button’s hit, the model is run, the widget outputs an Orange Data Table, we use the Line Chart Widget to visualize results of simulation:

Other widgets have been used to perform step by step execution of simulation, by simply running the results through As Timeseries while indicating Time as the index variable:

Next, using Time Slice which was adapted to commit a cumulative step by step simulation to the Line Chart:

In another example; we run the simulation in two different regions:

• Paris (France)

• Algiers(Algeria) and Casablanca (Morocco)

with different simulation parameters each time, we variate the total population, contact infectivity and the stocks (infectious, recovered, susceptible). We simply add the results together using Concatenate:

We can also show the results in Geo Map, after adding Latitude and Longitude information: