As used in the planning literature, an indicator is a variable that conveys information on the condition or trend of an attribute of the system considered. The indicator will then have a specific value at a given time. For UrbanSim, indicators provide the principal mechanism for presenting simulation results to modelers and other stakeholders so that they can be assessed and compared. In addition, modelers use indicators diagnostically to help assess whether the system is operating in a reasonable fashion and to help debug problems.
We will often be interested in the value of an indicator at different levels of aggregation, for example, population in each grid cell, in different political divisions of the region, and for the region as a whole. We will often also be interested in the change in the value of an indicator in successive years, or from each year of the simulation to the baseyear, or between two different scenarios. Indicator values should be displayed in an appropriate way, for example, using graphs, tables, or choropleth maps. Some key indicators for both policy evaluation and model diagnosis include population, residential units, land value, employment, and square feet of commercial, industrial, and governmental space, all at various levels of aggregation, from the grid cell up.
Requests for indicator visualizations can be made using either a graphical interface (Section 5.2), or a Python script (Section 5.3). The GUI is more user-friendly, but only allows a single indicator request to be made at a time. Scripts are Python code, but one script can specify an entire batch of indicators to be computed. Some of the more advanced functionality is also available only through the script.
There is online documentation for some of the indicators, linked from http://www.urbansim.org/docs/indicators/index.xml. See Section 11.6 for information on writing indicator documentation. (Formerly we computed the values for indicators using SQL queries, but this proved too slow in many cases, so we switched to using Opus attributes exclusively. There was much more extensive documentation for the SQL versions of the indicators; if there is demand for this and as time allows, we will also provide documentation for other indicators represented as Opus attributes.)