edm Description

edm implements a series of tools that can be used for empirical dynamic modeling in Stata. The command keyword is edm, and should be immediately followed by the subcommand explore or xmap. A dataset must be declared as time-series or panel data by the tsset or xtset command prior to using the edm command, and time-series operators including l., f., d., and s. can be used (the last for seasonal differencing).

Installation

To install the stable version directly through Stata:

ssc install edm, replace

To install the latest development version, first install the stable version from SSC then inside Stata run:

edm update, development replace

The source code for the package is available here.

Example Data and Do File

This paper describes the use of edm in Stata. The slides for the talk given at the QMNET seminar series is available here.

The Chicago crime dataset used in the edm paper can be downloaded here.

This do file contains the a series of examples of the edm command using both a synthetic dataset and the Chicago crime dataset.

Plugins

The edm command will support a plugin system which facilities some automated common analyses. This feature is under development and the files below provide some selected early stage prototypes allowing automated Simplex Projections, S-maps, Coprediction, and Convergent Cross Mapping as well as relevant hypothesis tests:

• edm Plugin - Automated time-series analysis (N=1 case)
• edm Plugin - Automated multi-spatial analysis (N>1 panel data case)
• edm Plugin - Automated multiple-edm analysis (N>1 panel data case)