Continuous Integration and Software Quality in Scientific Software: A Large-Scale Empirical Analysis of GitHub Repositories
DOI: 10.64898/2026.10.01.755028
|
Read the paper on bioRxiv
|
jakobilab.org
Repository Type
All
open_source
bioinformatics
astrophysics
software_engineering
image_recognition
Metric
All
ci_coverage
test_coverage
First Repo Type
open_source
bioinformatics
astrophysics
software_engineering
image_recognition
Second Repo Type
open_source
bioinformatics
astrophysics
software_engineering
image_recognition
Radar View
Top / Mid / Bottom 100 (Fig 1)
Average Profile by Domain
Top 20 vs Bottom 20 by Domain
Domain Scope
All Domains
Bioinformatics Only
Split by
None
Tests
CI
QA Bucket
Citation Plot
Mean vs RQI (loess)
Median vs RQI (loess)
Violin by RQI bin
Mean bar by RQI bin
Median bar (Low/High)
Median bar (RQI coarse bins)
Median bar (RQI fine bins)
Median bar, published only (Fig 5)
Citations vs CI
Median RQI by publication status
RQI distribution by publication status
RQI components by publication (Fig 4a)
Citations vs RQI + Age
Scatter + LM (published)
Funding Plot
Median RQI by Funding (Bioinformatics)
RQI Components by Funding, Fig 4b (Bioinformatics)
RQI vs Developer Award (Bioinformatics)
RQI vs Organization Award (Bioinformatics)
RQI Components by Funding (All Domains)
RQI vs Developer Award (All Domains)
RQI vs Organization Award (All Domains)
CI & Test Coverage
Distribution
ANOVA / Tukey
Pairwise Comparison
CI & Test Impact
Radar
Age Cohorts
Language
Activity / Survival
Citations
Funding
Summary Tables
CI and Testing Prevalence by Quality Tier
Prevalence Regression Across RQI (Fig 2)
Quality Distribution by Type (stacked)
Overall Distribution + Q-Q
Density by Type (faceted)
Bioinformatics vs All — Tukey Significant Pairs
Tukey HSD Pairwise Comparison
CI Impact
Test Impact
Repository Quality Radar Profiles
Mean RQI by Age Cohort — CI
Mean RQI by Age Cohort — Testing
Bioinformatics CI & Test Coverage by Cohort
Combined View (Fig 3)
Mean RQI by Primary Language
Odds Ratios: Predictors of Repo Activity
Activity Rate by CI/Test Presence and Domain
Table 1 — Mean Scores by Repository Type
Table S1 — Most Frequent Languages