This vignette demonstrates how to create a small project using
renv for dependency management. We will initialize a
project, install the psych package, and run a simple factor
analysis on one of the datasets that ships with that package, all while
using reproducibleRchunks.
First, inside your R project, initialize renv to track
and manage package dependencies:
renv creates a project-specific library and writes a
lock file to record package versions. Your R session will be restarted
(if running inside RStudio). Next we install the psych
package as well as the reproducibleRchunks in the managed
environment:
After installation, snapshot the project state so the lock file is updated:
With the packages installed and recorded, we can load
psych and run a quick factor analysis on the
bfi dataset that comes with that package. The following
code chunk could now be a part of an R Markdown document inside your
project:
library(psych)
#> Warning: Paket 'psych' wurde unter R Version 4.4.3 erstellt
# Use the first 200 cases of the bfi dataset for speed
bfi_subset <- head(bfi, 200)
fa_result <- fa(bfi_subset[, 1:10], nfactors = 3)
#> Lade nötigen Namensraum: GPArotation
fa_result
#> Factor Analysis using method = minres
#> Call: fa(r = bfi_subset[, 1:10], nfactors = 3)
#> Standardized loadings (pattern matrix) based upon correlation matrix
#> MR2 MR1 MR3 h2 u2 com
#> A1 0.01 -0.19 0.47 0.32 0.68 1.3
#> A2 0.10 0.51 -0.33 0.53 0.47 1.8
#> A3 0.00 0.83 0.00 0.69 0.31 1.0
#> A4 0.14 0.47 0.21 0.23 0.77 1.6
#> A5 -0.06 0.73 0.00 0.52 0.48 1.0
#> C1 0.65 -0.01 -0.05 0.43 0.57 1.0
#> C2 0.72 0.04 0.14 0.52 0.48 1.1
#> C3 0.66 -0.02 0.01 0.43 0.57 1.0
#> C4 -0.53 0.03 0.25 0.37 0.63 1.4
#> C5 -0.41 -0.02 0.12 0.20 0.80 1.2
#>
#> MR2 MR1 MR3
#> SS loadings 1.88 1.81 0.54
#> Proportion Var 0.19 0.18 0.05
#> Cumulative Var 0.19 0.37 0.42
#> Proportion Explained 0.44 0.43 0.13
#> Cumulative Proportion 0.44 0.87 1.00
#>
#> With factor correlations of
#> MR2 MR1 MR3
#> MR2 1.00 0.22 -0.13
#> MR1 0.22 1.00 -0.36
#> MR3 -0.13 -0.36 1.00
#>
#> Mean item complexity = 1.2
#> Test of the hypothesis that 3 factors are sufficient.
#>
#> df null model = 45 with the objective function = 2.35 with Chi Square = 457.46
#> df of the model are 18 and the objective function was 0.13
#>
#> The root mean square of the residuals (RMSR) is 0.03
#> The df corrected root mean square of the residuals is 0.05
#>
#> The harmonic n.obs is 199 with the empirical chi square 19.98 with prob < 0.33
#> The total n.obs was 200 with Likelihood Chi Square = 24.42 with prob < 0.14
#>
#> Tucker Lewis Index of factoring reliability = 0.961
#> RMSEA index = 0.042 and the 90 % confidence intervals are 0 0.081
#> BIC = -70.95
#> Fit based upon off diagonal values = 0.98
#> Measures of factor score adequacy
#> MR2 MR1 MR3
#> Correlation of (regression) scores with factors 0.88 0.90 0.69
#> Multiple R square of scores with factors 0.77 0.81 0.48
#> Minimum correlation of possible factor scores 0.54 0.63 -0.04Reproducibility Checks
✅bfi_subset: REPRODUCTION SUCCESSFUL
✅fa_result: REPRODUCTION SUCCESSFUL
This chunk is executed with reproducibleR, so
fingerprints of bfi_subset and fa_result are
stored for later reproducibility checks.
Note that if you operate inside RStudio, RStudio will ask your permission to install all packages necessary to render Markdown files inside your managed environment.
If you share your project, others can reproduce the same setup by
calling renv::restore() in the project directory. This will
install the recorded package versions and allow them to re-run the
analyses above with verified reproducibility.
If you host your project on GitHub, you can add a github action that automatically verifies reproduction in the GitHub cloud:
Depending on whether reproduction succeeded of failed, a badge is generated and added and committed to your repository. This is how the badges look like depending on the status:
or 