diff --git a/DESCRIPTION b/DESCRIPTION
index 583b255..e1a6646 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: mrgda
 Title: Tools for Data Assembly
-Version: 0.13.0
+Version: 0.13.0.9000
 Authors@R: 
     c(
     person(given = "Eric", family = "Anderson", email = "andersone@metrumrg.com", role = c("aut", "cre")),
@@ -16,6 +16,7 @@ Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.3.3
 Imports: 
+    bslib,
     cli (>= 3.3.0),
     dplyr,
     glue (>= 1.6.2),
@@ -36,7 +37,10 @@ Imports:
     stringr,
     knitr,
     readr,
-    tidyselect
+    shiny,
+    tidyselect,
+    ggplot2,
+    plotly
 Suggests: 
     testthat (>= 3.1.10),
     withr (>= 2.5.0),
diff --git a/NAMESPACE b/NAMESPACE
index f4ec63e..a1e343d 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -7,6 +7,7 @@ export(explain)
 export(query_src_list)
 export(read_csv_dots)
 export(read_src_dir)
+export(visualize_data)
 export(write_csv_dots)
 export(write_derived)
 import(rlang)
diff --git a/NEWS.md b/NEWS.md
index 9c74e7d..0d89eab 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -2,6 +2,7 @@
 
 ## New features and changes
 
+- Added `visualize_data` function to allow users to investigate data sets. (#253)
 - `query_src_list` now searches across values in addition to column and label names. (#251)
 
 # mrgda 0.13.0
diff --git a/R/app_exit_hint.R b/R/app_exit_hint.R
new file mode 100644
index 0000000..481d716
--- /dev/null
+++ b/R/app_exit_hint.R
@@ -0,0 +1,12 @@
+#' @noRd
+show_app_exit_hint <- function(app_name) {
+  cli::cli_rule(left = paste("Launching", app_name))
+  cli::cli_alert_info("{.strong How to close the app and free your R console:}")
+  cli::cli_ul()
+  cli::cli_li("EITHER: close the browser tab/window running {.emph {app_name}}")
+  cli::cli_li(
+    "OR: click the {.strong Stop} button (red stop sign) above the R Console in RStudio"
+  )
+  cli::cli_end()
+  invisible()
+}
diff --git a/R/visualize-data.R b/R/visualize-data.R
new file mode 100644
index 0000000..0c66816
--- /dev/null
+++ b/R/visualize-data.R
@@ -0,0 +1,385 @@
+#' Load the mrgda Shiny app
+#'
+#' Launch an interactive Shiny app for exploring a CSV data set. The Visualizer view
+#' provides a scatter plot with selectable X/Y axes, optional color grouping, hover
+#' fields based on the active axis/color/facet/filter selections, and ad hoc filters
+#' for numeric ranges or categorical values.
+#'
+#' CSV files are read with [read_csv_dots()], so "." values are interpreted as `NA`.
+#' If a YAML specification is supplied, it is loaded with [yspec::ys_load()] and
+#' applied via [yspec::ys_factors()] to coerce labeled factors before rendering.
+#'
+#' @param .csv_path Path to a CSV data file.
+#' @param .spec_path Optional path to a YAML specification file for factor metadata.
+#' @return A Shiny app object.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' visualize_data("analysis/adsl.csv")
+#' visualize_data("analysis/adsl.csv", "analysis/adsl.yaml")
+#' }
+visualize_data <- function(.csv_path, .spec_path = NULL) {
+  if (!file.exists(.csv_path)) {
+    stop("`.csv_path` does not exist: ", .csv_path, call. = FALSE)
+  }
+
+  data <- read_csv_dots(.csv_path)
+
+  # Apply factors to data
+  if (!is.null(.spec_path)) {
+
+    # Confirm spec path exists
+    if (!file.exists(.spec_path)) {
+      stop("`.spec_path` does not exist: ", .spec_path, call. = FALSE)
+    }
+
+    # Load in spec
+    spec <- yspec::ys_load(.spec_path)
+
+    data <- yspec::ys_add_factors(data, spec, .suffix = "")
+  }
+
+  data_vars <- names(data)
+  default_x <- if (length(data_vars) > 0) data_vars[[1]] else ""
+  default_y <- if (length(data_vars) > 1) data_vars[[2]] else default_x
+  default_hover <- unique(c(default_x, default_y))
+
+  ui <- bslib::page_fluid(
+    theme = bslib::bs_theme(version = 5),
+    shiny::tabsetPanel(
+      shiny::tabPanel(
+        title = "General visualizer",
+        bslib::layout_sidebar(
+          sidebar = bslib::sidebar(
+            shiny::selectInput(
+              inputId = "x_var",
+              label = "X-axis",
+              choices = data_vars,
+              selected = default_x
+            ),
+            shiny::selectInput(
+              inputId = "y_var",
+              label = "Y-axis",
+              choices = data_vars,
+              selected = default_y
+            ),
+            shiny::selectInput(
+              inputId = "color_var",
+              label = "Color by",
+              choices = c("None" = "", data_vars),
+              selected = ""
+            ),
+            shiny::selectInput(
+              inputId = "facet_var",
+              label = "Facet by",
+              choices = c("None" = "", data_vars),
+              selected = ""
+            ),
+            shiny::conditionalPanel(
+              condition = "input.facet_var !== ''",
+              shiny::checkboxInput(
+                inputId = "facet_free_scales",
+                label = "Free facet axes",
+                value = FALSE
+              )
+            ),
+            shiny::selectizeInput(
+              inputId = "hover_vars",
+              label = "Hover variables",
+              choices = data_vars,
+              selected = default_hover,
+              multiple = TRUE
+            ),
+            shiny::selectizeInput(
+              inputId = "filter_vars",
+              label = "Filter variables",
+              choices = data_vars,
+              multiple = TRUE
+            ),
+            shiny::uiOutput("filter_ui")
+          ),
+          shiny::uiOutput("scatter_plot_ui")
+        )
+      )
+    )
+  )
+
+  server <- function(input, output, session) {
+    session$onSessionEnded(function() shiny::stopApp())
+
+    # show hint once the server has started
+    show_app_exit_hint("diffDashboard")
+
+    filter_map <- shiny::reactive({
+      vars <- input$filter_vars
+      if (is.null(vars) || length(vars) == 0) {
+        return(NULL)
+      }
+      ids <- make.names(vars, unique = TRUE)
+      stats::setNames(ids, vars)
+    })
+
+    output$filter_ui <- shiny::renderUI({
+      vars <- input$filter_vars
+      if (is.null(vars) || length(vars) == 0) {
+        return(NULL)
+      }
+      mapping <- filter_map()
+      ui_list <- lapply(vars, function(var) {
+        id_base <- mapping[[var]]
+        filter_data <- data[[var]]
+        if (is.numeric(filter_data)) {
+          rng <- range(filter_data, na.rm = TRUE)
+          if (!all(is.finite(rng))) {
+            return(NULL)
+          }
+          shiny::sliderInput(
+            inputId = paste0("filter_num__", id_base),
+            label = paste0(var, " range"),
+            min = rng[1],
+            max = rng[2],
+            value = rng
+          )
+        } else {
+          choices <- sort(unique(as.character(filter_data)))
+          shiny::selectizeInput(
+            inputId = paste0("filter_cat__", id_base),
+            label = paste0(var, " values"),
+            choices = choices,
+            selected = choices,
+            multiple = TRUE
+          )
+        }
+      })
+      do.call(shiny::tagList, ui_list)
+    })
+
+    auto_hover_vars <- shiny::reactive({
+      vars <- c(
+        input$x_var,
+        input$y_var,
+        if (!is.null(input$color_var) && nzchar(input$color_var)) input$color_var else NULL,
+        if (!is.null(input$facet_var) && nzchar(input$facet_var)) input$facet_var else NULL,
+        input$filter_vars
+      )
+      vars <- vars[!is.na(vars) & nzchar(vars)]
+      unique(vars)
+    })
+
+    shiny::observeEvent(
+      list(
+        input$x_var,
+        input$y_var,
+        input$color_var,
+        input$facet_var,
+        input$filter_vars
+      ),
+      {
+        selected <- unique(c(auto_hover_vars(), input$hover_vars))
+        shiny::updateSelectizeInput(
+          session,
+          "hover_vars",
+          choices = data_vars,
+          selected = selected
+        )
+      },
+      ignoreInit = FALSE
+    )
+
+    filtered_data <- shiny::reactive({
+      vars <- input$filter_vars
+      if (is.null(vars) || length(vars) == 0) {
+        return(data)
+      }
+      mapping <- filter_map()
+      plot_data <- data
+      for (var in vars) {
+        id_base <- mapping[[var]]
+        filter_data <- plot_data[[var]]
+        if (is.numeric(filter_data)) {
+          range_id <- paste0("filter_num__", id_base)
+          rng <- input[[range_id]]
+          if (is.null(rng) || length(rng) != 2) {
+            next
+          }
+          keep <- filter_data >= rng[1] & filter_data <= rng[2]
+          plot_data <- plot_data[keep, , drop = FALSE]
+        } else {
+          cat_id <- paste0("filter_cat__", id_base)
+          vals <- input[[cat_id]]
+          if (is.null(vals) || length(vals) == 0) {
+            plot_data <- plot_data[FALSE, , drop = FALSE]
+            next
+          }
+          keep <- as.character(filter_data) %in% vals
+          plot_data <- plot_data[keep, , drop = FALSE]
+        }
+      }
+      plot_data
+    })
+
+    plot_spec <- shiny::reactive({
+      plot_data <- filtered_data()
+      shiny::req(input$x_var, input$y_var)
+      x <- plot_data[[input$x_var]]
+      y <- plot_data[[input$y_var]]
+      keep_xy <- !is.na(x) & !is.na(y)
+      plot_data <- plot_data[keep_xy, , drop = FALSE]
+      x <- x[keep_xy]
+      y <- y[keep_xy]
+      if (nrow(plot_data) == 0) {
+        return(list(empty = TRUE))
+      }
+      hover_vars <- unique(c(auto_hover_vars(), input$hover_vars))
+      hover_vars <- hover_vars[hover_vars %in% data_vars]
+      hover_text <- NULL
+      if (length(hover_vars) > 0) {
+        hover_df <- plot_data[, hover_vars, drop = FALSE]
+        hover_df[] <- lapply(hover_df, function(col) {
+          if (is.factor(col)) {
+            as.character(col)
+          } else {
+            col
+          }
+        })
+        hover_text <- apply(
+          hover_df,
+          1,
+          function(row) paste(paste0(hover_vars, ": ", row), collapse = "<br>")
+        )
+      }
+
+      color_data <- NULL
+      if (!is.null(input$color_var) && nzchar(input$color_var)) {
+        color_data <- plot_data[[input$color_var]]
+      }
+
+      facet_data <- NULL
+      if (!is.null(input$facet_var) && nzchar(input$facet_var)) {
+        facet_data <- plot_data[[input$facet_var]]
+      }
+
+      base_df <- data.frame(
+        x = x,
+        y = y,
+        hover_text = if (is.null(hover_text)) NA_character_ else hover_text,
+        stringsAsFactors = FALSE
+      )
+      if (!is.null(color_data)) {
+        base_df$color <- if (is.factor(color_data)) as.character(color_data) else color_data
+      }
+      if (!is.null(facet_data)) {
+        facet_vals <- if (is.factor(facet_data)) as.character(facet_data) else facet_data
+        facet_vals[is.na(facet_vals)] <- "(Missing)"
+        base_df$facet <- facet_vals
+      }
+
+      list(
+        empty = FALSE,
+        base_df = base_df,
+        has_color = !is.null(color_data),
+        has_facet = !is.null(facet_data),
+        tooltip = if (is.null(hover_text)) "x+y" else "text",
+        x_label = input$x_var,
+        y_label = input$y_var,
+        color_label = if (!is.null(color_data)) input$color_var else NULL,
+        facet_free_scales = isTRUE(input$facet_free_scales)
+      )
+    })
+
+    build_plot <- function(plot_df, spec, facets_per_plot = NULL) {
+      gg_args <- list(
+        data = plot_df,
+        mapping = ggplot2::aes(
+          x = .data$x,
+          y = .data$y,
+          text = .data$hover_text
+        )
+      )
+      if (isTRUE(spec$has_color)) {
+        gg_args$mapping <- ggplot2::aes(
+          x = .data$x,
+          y = .data$y,
+          text = .data$hover_text,
+          color = .data$color
+        )
+      }
+
+      gg_plot <- do.call(ggplot2::ggplot, gg_args) +
+        ggplot2::geom_point() +
+        ggplot2::labs(
+          x = spec$x_label,
+          y = spec$y_label,
+          color = spec$color_label
+        ) +
+        ggplot2::theme_bw() +
+        ggplot2::theme(
+          panel.grid.minor = ggplot2::element_blank(),
+          panel.grid.major = ggplot2::element_line(color = "grey85"),
+          panel.background = ggplot2::element_rect(fill = "white"),
+          plot.background = ggplot2::element_rect(fill = "white")
+        )
+
+      if (isTRUE(spec$has_facet)) {
+        facet_scales <- if (isTRUE(spec$facet_free_scales)) "free" else "fixed"
+        gg_plot <- gg_plot + ggplot2::facet_wrap(~facet, ncol = 3, scales = facet_scales)
+      }
+
+      plotly::ggplotly(gg_plot, tooltip = spec$tooltip)
+    }
+
+    output$scatter_plot_ui <- shiny::renderUI({
+      spec <- plot_spec()
+      if (isTRUE(spec$empty)) {
+        return(
+          shiny::div(
+            class = "text-muted",
+            "No rows remain after filtering for non-missing X/Y values."
+          )
+        )
+      }
+      if (!isTRUE(spec$has_facet)) {
+        output$scatter_plot_1 <- plotly::renderPlotly({
+          build_plot(spec$base_df, spec)
+        })
+        return(plotly::plotlyOutput("scatter_plot_1", height = "800px"))
+      }
+
+      facet_levels <- unique(spec$base_df$facet)
+      facet_groups <- split(
+        facet_levels,
+        ceiling(seq_along(facet_levels) / 9)
+      )
+
+      ui_list <- lapply(seq_along(facet_groups), function(i) {
+        output_id <- paste0("scatter_plot_", i)
+        facet_subset <- facet_groups[[i]]
+        output[[output_id]] <- plotly::renderPlotly({
+          plot_df <- spec$base_df[spec$base_df$facet %in% facet_subset, , drop = FALSE]
+          build_plot(plot_df, spec)
+        })
+        plotly::plotlyOutput(output_id, height = "800px")
+      })
+
+      if (length(ui_list) > 1) {
+        tabs <- lapply(seq_along(ui_list), function(i) {
+          shiny::tabPanel(
+            title = paste0("Facets ", i),
+            ui_list[[i]]
+          )
+        })
+        do.call(shiny::tabsetPanel, tabs)
+      } else {
+        do.call(shiny::tagList, ui_list)
+      }
+    })
+  }
+
+  shiny::shinyApp(
+    ui = ui,
+    server = server,
+    options = list(.csv_path = .csv_path, .spec_path = .spec_path,
+                   launch.browser = TRUE, quiet = TRUE)
+  )
+}
diff --git a/_pkgdown.yml b/_pkgdown.yml
index 0528b02..bb41a48 100644
--- a/_pkgdown.yml
+++ b/_pkgdown.yml
@@ -28,3 +28,4 @@ reference:
     - explain
     - read_csv_dots
     - write_csv_dots
+    - visualize_data
diff --git a/man/visualize_data.Rd b/man/visualize_data.Rd
new file mode 100644
index 0000000..a1281ee
--- /dev/null
+++ b/man/visualize_data.Rd
@@ -0,0 +1,33 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/visualize-data.R
+\name{visualize_data}
+\alias{visualize_data}
+\title{Load the mrgda Shiny app}
+\usage{
+visualize_data(.csv_path, .spec_path = NULL)
+}
+\arguments{
+\item{.csv_path}{Path to a CSV data file.}
+
+\item{.spec_path}{Optional path to a YAML specification file for factor metadata.}
+}
+\value{
+A Shiny app object.
+}
+\description{
+Launch an interactive Shiny app for exploring a CSV data set. The Visualizer view
+provides a scatter plot with selectable X/Y axes, optional color grouping, hover
+fields based on the active axis/color/facet/filter selections, and ad hoc filters
+for numeric ranges or categorical values.
+}
+\details{
+CSV files are read with \code{\link[=read_csv_dots]{read_csv_dots()}}, so "." values are interpreted as \code{NA}.
+If a YAML specification is supplied, it is loaded with \code{\link[yspec:ys_load]{yspec::ys_load()}} and
+applied via \code{\link[yspec:ys_factors]{yspec::ys_factors()}} to coerce labeled factors before rendering.
+}
+\examples{
+\dontrun{
+visualize_data("analysis/adsl.csv")
+visualize_data("analysis/adsl.csv", "analysis/adsl.yaml")
+}
+}
diff --git a/tests/testthat/test-visualize-data.R b/tests/testthat/test-visualize-data.R
new file mode 100644
index 0000000..2332f7a
--- /dev/null
+++ b/tests/testthat/test-visualize-data.R
@@ -0,0 +1,29 @@
+test_that("visualize_data validates input paths", {
+
+  expect_error(
+    visualize_data("does-not-exist.csv"),
+    "`.csv_path` does not exist"
+  )
+
+  csv_path <- withr::local_tempfile(fileext = ".csv")
+  writeLines(c("a,b", "1,2"), csv_path)
+
+  expect_error(
+    visualize_data(csv_path, "missing-spec.yaml"),
+    "`.spec_path` does not exist"
+  )
+})
+
+test_that("visualize_data returns a shiny app with options", {
+
+  csv_path <- withr::local_tempfile(fileext = ".csv")
+  writeLines(c("a,b", "1,2"), csv_path)
+
+  app <- visualize_data(csv_path)
+
+  expect_true(inherits(app, "shiny.appobj"))
+  expect_equal(app$options$.csv_path, csv_path)
+  expect_null(app$options$.spec_path)
+  expect_true(isTRUE(app$options$launch.browser))
+  expect_true(isTRUE(app$options$quiet))
+})