We are delighted to introduce you to Prest 1.0.
Prest is a stand-alone, user-friendly and open-source desktop application for computational revealed preference analysis that we published in October 2018 in the Journal of Open Source Software.
It operationalizes and extends classic and recent ideas from the literature, enabling users to analyse choice datasets of the kind that economists, psychologists and consumer/marketing researchers often generate through experiments or surveys.
Prest can be used for both rational and behavioral revealed preference analysis. Although budgetary choice datasets are also supported, the key novelties of this program currently pertain to general datasets where choices are made from finite menus of discrete alternatives. For example, using combinatorial-optimization methods, Prest allows for estimating the sampled decision makers’ preferences from such general datasets by finding out “how close” their observed choices are to being perfectly explainable by utility maximization *or* by one of several models of bounded-rational decision making that generalize utility maximization.
Prest 1.0 binaries are available for Windows & macOS and can be downloaded from https://prestsoftware.com
. (The macOS binary is currently unsigned and thus its first run has to be confirmed in System Preferences/Security and Privacy)
The above website also contains detailed documentation for Prest’s features, and several example datasets.
The core of Prest is written in Rust for stability, speed and ease of parallelisation, while the graphical user interface is written in Python. It is our aim to continue working towards improving and expanding the software’s functionality in the future. Although we regret that we cannot offer technical support at present, we welcome issue reports or even pull requests on the source code repository: https://github.com/prestsoftware/prest
We conclude with the project’s Declarations:
1. Prest is open-source software and its latest version will always be available online for free.
2. Prest does not collect any data entered by its users.
Georgios Gerasimou & Matúš Tejiščák