This Shiny app accompanies the study titled "Applying an interpretable machine learning approach to assess intraspecific trait variation under landscape-scale population differentiation," authored by Sambadi Majumder and Dr. Chase Mason. The study, accessible through this DOI link, investigates the functional trait data of Helianthus annuus genotypes from the HeliantHome database.
The functional trait data is from the HeliantHome database, publicized in Bercovich et al., 2022 (DOI link). Genotypes were selected based on their occurrence within the Level I ecoregions of the Great Plains and North American Deserts, correlated with ecoregion shapefiles from the U.S. Environmental Protection Agency.
- Programming Language: R
- Key Packages:
shiny
for app functionality.leaflet
for interactive mapping 🗺️.ggplot2
andplotly
for creating visualizations 📊.dplyr
for data manipulation 🔨.sf
for handling spatial data 🌍.
Displays a map of Helianthus annuus populations' distribution within the Great Plains and North American Deserts ecoregions.
Reveals traits exhibiting divergence between the Desert and Plains populations, highlighting those most predictive of ecoregion.
Shows accumulated local effects plots articulating the impact of each divergent trait on ecoregion classification.
Demonstrates an interpretable machine learning approach for identifying ecoregion-predictive traits, essential for ecological strategy research in Helianthus annuus.
Navigate between study components through an intuitive interface, engaging with visualizations for a deeper understanding of the findings.
Explore the interactive visualizations on the Shiny app here.