Prest

Preference estimation from choice data: https://prestsoftware.com

Reach us at contact AT prestsoftware DOT com.

Introduction

Prest is a free and user-friendly desktop application for computational revealed preference analysis. It allows for processing choice datasets that economists, psychologists and consumer/marketing researchers often generate through experiments, market studies or surveys.

Documentation and downloads

Pre-built Prest binaries can be downloaded from the landing page of the documentation at https://prestsoftware.com.

Building and running Prest

Dependencies:

Rust 1.26 stable + Cargo
Python 3 (we use 3.6 and 3.7) + pip

Install the dependencies, compile everything and run:

$ pip install --user -r gui/requirements.txt
$ make run

Optionally, work in a virtual environment:

$ python3 -m venv prest.env
$ source prest.env/bin/activate
# The next line is different from the previous installation command.
$ pip install -r gui/requirements.txt
$ make run

The build invoked by the commands above will, among other things, build the HTML documentation, which is required for the help features of Prest. It will also typecheck the code using mypy.

Testing

$ make test      # quick test during development
$ make fulltest  # includes long-running tests

The most comprehensive test is the integration test, which runs the whole pipeline including the Rust core on the example datasets. It is invoked in the course of the above commands; make test uses only the small example datasets, while make fulltest uses all of them.

Packaging

We build stand-alone binaries using PyInstaller. These build scripts are not published at the moment.

License

Prest consists of two separate programs: GUI (GNU GPL) and core (BSD-3-Clause). The full license texts can be found in the corresponding subdirectories of the source code.

Citation

If you would like to cite Prest in your academic work, you can use the following reference:

Georgios Gerasimou and Matúš Tejiščák (2018) Prest: Open-Source Software for Computational Revealed Preference Analysis, Journal of Open Source Software, 3(30), 1015, doi:10.21105.joss.01015.

Declarations

Prest does not collect any data entered by its users.

The latest version of Prest will always be available online for free.

DavideNardone/prest