Marimo dashboard proof-of-concept

.gitignore

@@ -1,4 +1,6 @@
 data/
+dashboard/public/data
+_site/
 
 ## Python.gitignore
 # https://github.com/github/gitignore/blob/5445a270254a0b4443b20ed0033c4094959f937e/Python.gitignore

dashboard/index.py

@@ -0,0 +1,97 @@
+import marimo
+
+__generated_with = "0.14.17"
+app = marimo.App(width="medium", app_title="EAVS Data Explorer")
+
+
+@app.cell
+def _(mo):
+    mo.md(
+        r"""
+    # EAVS Data Explorer
+
+    This is a proof-of-concept using Marimo to export a deployable dashboard. It generates static HTML with WASM for data-loading and interactivity, and can be deployed with a static site host like GitHub pages. 
+    """
+    )
+    return
+
+
+@app.cell
+def _():
+    import marimo as mo
+    import polars as pl
+    import plotly.express as px
+    return mo, pl, px
+
+
+@app.cell
+def _(mo):
+    data_path = mo.notebook_location() / "public" / "data" / "combined.csv"
+    return (data_path,)
+
+
+@app.cell
+def _(data_path, pl):
+    df = pl.read_csv(str(data_path))
+    return (df,)
+
+
+@app.cell
+def _():
+    return
+
+
+@app.cell
+def _(df):
+    dimensions = (
+        "year",
+        "fips_code",
+        "jurisdiction_name",
+        "state",
+        "state_abbr",
+    )
+
+    measures = tuple(
+        col for col, dtype in zip(df.columns, df.dtypes)
+        if dtype.is_numeric() and col not in dimensions
+    )
+
+    years = df.select("year").unique().to_series().to_list()
+    return dimensions, measures, years
+
+
+@app.cell
+def _(measures, mo, years):
+    year_dropdown = mo.ui.dropdown(options=years, label="Year", value=years[0])
+    yvar_dropdown = mo.ui.dropdown(options=measures, label="Y Variable", value=measures[0])
+    xvar_dropdown = mo.ui.dropdown(options=measures, label="X Variable", value=measures[1])
+    controls = mo.vstack([year_dropdown, yvar_dropdown, xvar_dropdown], justify="center")
+    return controls, xvar_dropdown, year_dropdown, yvar_dropdown
+
+
+@app.cell
+def _(df, dimensions, pl, px, xvar_dropdown, year_dropdown, yvar_dropdown):
+    plot = px.scatter(
+        df.filter(pl.col("year")==year_dropdown.value), 
+        x=xvar_dropdown.value, 
+        y=yvar_dropdown.value, 
+        hover_data=dimensions
+    )
+
+    return (plot,)
+
+
+@app.cell
+def _(controls, mo, plot):
+    mo.hstack([controls, plot])
+    return
+
+
+@app.cell
+def _(df, mo):
+    mo.ui.dataframe(df)
+    return
+
+
+if __name__ == "__main__":
+    app.run()

justfile

@@ -15,3 +15,8 @@ lint:
 format:
     ruff format
     ruff check --fix
+
+# Export dashboard
+export-dashboard:
+    cp data/cleaned/combined.csv dashboard/public/data/combined.csv
+    marimo export html-wasm dashboard/index.py -o _site --mode run

pyproject.toml

@@ -15,8 +15,11 @@ requires-python = "~=3.12.0"
 dependencies = [
     "httpx",
     "loguru",
+    "marimo",
     "pandas",
     "pandera[io]",
+    "plotly",
+    "polars",
     "python-calamine",
     "python-dotenv",
     "pyyaml",