TL;DR
This paper introduces Pabulib, a new online library for participatory budgeting (PB) data, and a standardized .pb file format for representing PB instances. This initiative aims to foster research in computational social choice applied to PB by providing accessible, structured datasets.
What Problem Does This Paper Address?
Participatory budgeting (PB) is a democratic process where citizens decide how to allocate public funds. While PB has gained global traction, research in applying computational social choice (CSC) methods to PB has been hampered by a lack of standardized, publicly available datasets. This makes it difficult for researchers to compare results, reproduce studies, and develop robust aggregation mechanisms for real-world PB scenarios. The paper addresses this by creating a centralized data repository and defining a clear, machine-readable format for PB data.
Methods
The core contribution of this paper is the definition of a .pb file format for representing participatory budgeting instances. This format is designed to be stored in a single UTF-8 text file and is structured into three distinct sections:
- META: Contains general metadata about the PB instance, such as the country, the governing unit (municipality, region, etc.), the specific instance (e.g., year), the total number of projects and votes, the budget amount, and the voting rule used.
- PROJECTS: Details each project, including a unique
project_id, itscost, and optional fields likename,category, andtargetdemographics. - VOTES: Encompasses the voter preferences. It includes
voter_idand optional demographic information likeageandsex. Thevoteitself is structured according to thevote_typespecified in the META section, supporting approval, ordinal, cumulative, and scoring vote types.
The paper also describes the Pabulib website (http://pabulib.org/) as the platform hosting these datasets.
Datasets
The paper introduces the Pabulib library as a collection of participatory budgeting datasets. While the paper itself does not present new datasets for training or evaluation, it establishes a format for future data contributions. The example provided in Section 3 illustrates the structure of a .pb file for a municipal PB instance in Wieliczka, Poland.
Key Insights and Ideas
- Standardization is Crucial for PB Research: The lack of a common data format has been a significant bottleneck for advancing research in PB, especially from a CSC perspective. Pabulib and its
.pbformat directly address this by providing a unified way to represent PB scenarios. - Flexibility in Vote Representation: The
.pbformat is designed to accommodate various voting mechanisms (approval, ordinal, cumulative, scoring), reflecting the diverse ways PB can be implemented in practice. - Rich Metadata for Context: The inclusion of a comprehensive META section allows for detailed contextualization of each PB instance, which is vital for understanding the nuances of different PB processes and for conducting more insightful research.
- Bridging Theory and Practice: By creating a publicly accessible library of real-world PB data, Pabulib aims to facilitate the application and testing of theoretical CSC models in practical, real-world settings.
Key Results
The primary result of this paper is the definition and introduction of the Pabulib library and its associated .pb file format. The paper provides a detailed specification of this format, outlining the mandatory and optional fields for each section (META, PROJECTS, VOTES) and how different vote types are encoded. The paper also makes the Pabulib website publicly accessible, serving as a central hub for PB data.
Limitations and Future Work
- Limited Data at Launch: While Pabulib is introduced as a library, the paper focuses on defining the format rather than presenting a large collection of diverse datasets. The success of Pabulib will depend on future contributions from researchers and practitioners.
- Future Rule Definitions: The paper mentions that "other rules will be defined in future versions" for the
rulefield in the META section, indicating that the current specification might be extended to support more aggregation rules. - Non-Standard Fields: The format allows for "non-standard fields" in both PROJECTS and VOTES sections, which offers flexibility but could also lead to data heterogeneity if not managed carefully.
- Encouraging Contributions: The authors express hope that Pabulib will foster meaningful research and encourage the community to contribute data, which is a key aspect of its future development.
Summary updated