About This Project
Understanding the landscape of academic conferences in systems and networks
Global Analysis
Comprehensive data from major systems conferences
This visualization analyzes data from top-tier systems and networks conferences including OSDI, ASPLOS, NSDI, SIGCOMM, EuroSys, ATC, and others. We track papers, committee composition, and geographical distributions to understand trends in academic research.
Historical Data
Tracking changes over two decades
Our dataset covers the period from 2000 to 2024, allowing for long-term trend analysis. We examine how the academic landscape has evolved, particularly the growth of contributions from Asia and the increasing involvement of major technology companies.
Key Metrics
What we measure and why it matters
- • Continental distribution of accepted papers
- • Evolution of Asian academic contributions
- • Big Tech vs Academia participation
- • Program Committee vs Papers geographic comparison
- • Diversity indices across conferences and committees
Data Sources
Quality and methodology
Our crawler gathers information from multiple sources: DBLP and Semantic ScholarAPIs provide structured metadata for papers (titles, authors, publication years, affiliations, and institutions). For some conferences where API data was incomplete or unavailable, data was manually crawled from official conference websites. Program committee member data were collected manually due to the heterogeneity of conference websites. The dataset is publicly available at github.com/Marina-LA/ConferenceData.
Methodology
Data Collection
Our dataset includes approximately 9,712 accepted papers and 14,996 non-deduplicated program committee members (6,917 unique individuals) across all conferences under study. Paper and author data were primarily collected automatically via our crawler using DBLP and Semantic Scholar APIs. For some conferences where API data was incomplete or unavailable, data was manually crawled from official conference websites. Program committee member data were manually compiled due to inconsistent website structures across conferences and years. Citation data were removed from the analysis after verification revealed significant inconsistencies in the API-provided data.
Continental Classification
Both papers and program committee members are mapped to continents based on author or member affiliations. If most authors belong to the same continent, the paper (or member) is assigned accordingly. In cases of ties, assignment is made uniformly at random. If most authors lack institutional information, assignment is based on the majority continent among remaining authors. If no country information is available, the entity remains unassigned. Over 90% of papers were assigned unambiguously to a single continent.
Big Tech vs Academia
We identify papers with authors affiliated with major technology companies (Google, Microsoft, Meta, Amazon, Apple, etc.) versus purely academic institutions.
Diversity Index
The Gini-Simpson Index measures diversity from 0 (homogeneous) to 1 (highly diverse). Higher values indicate more balanced distribution across categories.
Committee vs Papers Gap Analysis
We calculate the representation gap as Committee % minus Papers % for each continent. A positive gap indicates committee over-representation, while a negative gap shows paper over-representation from that region.
Data Quality
Missing affiliation data were retrieved manually from official conference websites when possible. Approximately 5% of papers in each conference were assigned to a continent despite most authors lacking affiliation data. While minor imperfections may exist, careful verification ensures they do not materially affect the qualitative conclusions of the study.
Authors
This project was developed by:
- • Pedro Garcia Lopez
- • Marina López Alet
- • Usama Benabdelkrim Zakan
- • Anwitaman Datta
Related Publication
Additional Resources
Access additional resources and data: