A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.
Python artificial-intelligencedagdata-sciencedata-visualizationdataflowdeveloper-toolsmachine-learningnotebookspipelinepythonreactivesqlweb-app
# /// script
# requires-python = ">=3.11"
# dependencies = [
# "marimo",
# "pandas==2.2.3",
# "matplotlib==3.9.2",
# "altair==5.4.1",
# ]
# ///
import marimo
__generated_with = "0.19.7"
app = marimo.App(width="full")
@app.cell(hide_code=True)
def _(mo):
mo.md("""
# Data Explorer
""")
return
@app.cell(hide_code=True)
def _(mo):
sample = "https://github.com/vega/vega/blob/main/docs/data/stocks.csv"
mo.md(
f"""
This notebook lets you upload a CSV and plot its columns.
You can download a <a href="{sample}" target="_blank">sample CSV</a> if you'd like.
"""
)
return
@app.cell(hide_code=True)
def _(mo, uploaded_file):
mo.md(
f"""
{mo.hstack([mo.md("**Upload a CSV.**")], justify="center")}
{uploaded_file}
"""
)
return
@app.cell
def _(io, mo, pd, uploaded_file):
mo.stop(not uploaded_file.name())
df = pd.read_csv(io.StringIO(uploaded_file.contents().decode()))
return (df,)
@app.cell
def _(df, mo):
mo.ui.table(df, page_size=5, selection=None)
return
@app.cell
def _(df, mo):
plot_type = mo.ui.dropdown(
["line", "hist"], value="line", label="Choose a plot type: "
)
x_column = mo.ui.dropdown(df.columns, label="Choose x-axis: ")
y_column = mo.ui.dropdown(df.columns, label="Choose y-axis: ")
color_column = mo.ui.dropdown(df.columns, label="Choose color-axis: ")
return color_column, plot_type, x_column, y_column
@app.cell
def _(color_column, mo, plot_type, x_column, y_column):
mo.hstack(
[x_column, y_column, color_column, plot_type], justify="space-around"
).callout(kind="warn" if not x_column.value else "neutral")
return
@app.cell
def _(alt, color_column, df, mo, plot_type, x_column, y_column):
mo.stop(not x_column.value)
def plot(x_column, y_column, color_column):
y_column = y_column or "count()"
title = f"{y_column} by {x_column}"
encoding = {"x": x_column, "y": y_column}
if color_column:
encoding["color"] = color_column
if plot_type.value == "line":
chart = alt.Chart(df).mark_line()
else:
chart = alt.Chart(df).mark_bar().encode(x=alt.X(x_column, bin=True))
return chart.encode(**encoding).properties(title=title, width="container")
plot(x_column.value, y_column.value, color_column.value)
return
@app.cell
def _(mo):
uploaded_file = mo.ui.file(filetypes=[".csv"], kind="area")
return (uploaded_file,)
@app.cell
def _():
import marimo as mo
import altair as alt
import io
import matplotlib.pyplot as plt
import pandas as pd
return alt, io, mo, pd
if __name__ == "__main__":
app.run()