projectLSA - Shiny Application for Latent Structure Analysis with a Graphical
User Interface
Provides an interactive Shiny-based toolkit for conducting
latent structure analyses, including Latent Profile Analysis
(LPA), Latent Class Analysis (LCA), Latent Trait Analysis
(LTA/IRT), Exploratory Factor Analysis (EFA), Confirmatory
Factor Analysis (CFA), and Structural Equation Modeling (SEM).
The implementation is grounded in established methodological
frameworks: LPA is supported through 'tidyLPA' (Rosenberg et
al., 2018) <doi:10.21105/joss.00978>, LCA through 'poLCA'
(Linzer & Lewis, 2011) <doi:10.32614/CRAN.package.poLCA> &
'glca' (Kim & Kim, 2024) <doi:10.32614/CRAN.package.glca>,
LTA/IRT via 'mirt' (Chalmers, 2012)
<doi:10.18637/jss.v048.i06>, and EFA via 'psych' (Revelle,
2025). SEM and CFA functionalities build upon the 'lavaan'
framework (Rosseel, 2012) <doi:10.18637/jss.v048.i02>. The
CFA/SEM module additionally supports multi-group invariance
testing, latent growth modelling, ordinal indicators estimated
from polychoric correlations, modification indices, and path
diagram visualisation. Every module can save and restore an
analysis session and export an 'R Markdown' HTML report whose
sections can be drafted with an optional AI assistant. The
assistant interprets the current results through one or several
large language models reached with user-supplied API keys,
including providers with a free tier. The application runs in a
responsive HTML interface. Users can upload datasets or use
built-in examples, fit models, compare fit indices, visualize
results, and export outputs without programming.
