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 = [
# "pandas==2.2.3",
# "vega-datasets==0.9.0",
# "requests==2.32.3",
# "altair==5.4.1",
# "marimo",
# ]
# ///
import marimo
__generated_with = "0.19.7"
app = marimo.App(width="full")
@app.cell
def _(chart, iss_df, mo, n_points_slider, refresh_interval_slider, refresher):
mo.hstack(
[
mo.vstack(
[
mo.md(
"## Settings | [`marimo.ui.slider`](https://docs.marimo.io/api/inputs/slider.html), [`marimo.ui.refresh`](https://docs.marimo.io/recipes.html#run-a-cell-on-a-timer)\n---"
),
refresh_interval_slider,
n_points_slider,
refresher,
mo.md(
"## ISS Positions | [`marimo.ui.altair_chart`](https://docs.marimo.io/api/plotting.html#marimo.ui.altair_chart)\n---"
),
mo.as_html(chart).style({"width": "700px"}),
],
align="center",
),
mo.vstack(
[
mo.md(
"## Data | [`marimo.as_html`](https://docs.marimo.io/api/html.html)`(pd.DataFrame)`\n---"
),
mo.as_html(iss_df),
]
),
],
justify="center",
wrap=True,
gap=3,
)
return
@app.cell
def _(alt, get_iss_positions, sphere, world):
hover = alt.selection_point(on="mouseover", clear="mouseout")
# iss positions
iss_df = get_iss_positions()
iss = (
alt.Chart(iss_df[["longitude", "latitude", "timestamp"]])
.mark_circle(
stroke="black",
size=100,
)
.encode(
longitude=alt.Longitude("longitude:Q"),
latitude="latitude:Q",
fill=alt.Fill(
"timestamp:Q", scale=alt.Scale(scheme="purples"), legend=None
),
strokeWidth=alt.condition(
hover, alt.value(3, empty=False), alt.value(0)
),
tooltip=[
alt.Tooltip("longitude:Q", title="Longitude", format=".4f"),
alt.Tooltip("latitude:Q", title="Latitude", format=".4f"),
alt.Tooltip(
"timestamp:T", title="Timestamp", format="%Y-%m-%d %H:%M:%S"
),
],
)
.add_params(hover)
)
chart = (
alt.layer(sphere, world, iss)
.project(type="naturalEarth1")
.properties(width=640, title="")
)
return chart, iss_df
@app.cell
def _(alt, data):
# load geo data from Vega Datasets
countries = alt.topo_feature(data.world_110m.url, 'countries')
# world base
sphere = alt.Chart(alt.sphere()).mark_geoshape(
fill="aliceblue", stroke="black", strokeWidth=1.5
)
# world map
world = alt.Chart(countries).mark_geoshape(
fill="mintcream", stroke="black", strokeWidth=0.35
)
return sphere, world
@app.cell
def _(n_points_slider, pd, refresh_interval_slider, refresher, requests, time):
def get_iss_positions(refresher=refresher):
refresher
timepoints = [int(time())]
while len(timepoints) <= n_points_slider.value:
timepoints.append(timepoints[-1] - refresh_interval_slider.value)
else:
timepoints.pop(0)
timepoints_str = str(timepoints)[1:-1].replace(" ", "")
iss_url = f"https://api.wheretheiss.at/v1/satellites/25544/positions?timestamps={timepoints_str}"
response = requests.get(iss_url)
df = pd.DataFrame(response.json())
df['timestamp'] = pd.to_datetime(df.timestamp, unit='s')
return df[['timestamp','latitude','longitude','altitude','velocity','visibility']]
return (get_iss_positions,)
@app.cell
def _(mo, refresh_interval_slider):
refresher = mo.ui.refresh(default_interval=f"{refresh_interval_slider.value}s")
return (refresher,)
@app.cell
def _(mo):
refresh_interval_slider = mo.ui.slider(start=5, stop=60, step=1, value=10, label="refresh interval (default = 10 sec)")
n_points_slider = mo.ui.slider(start=5, stop=30, step=1, value=15, label="number of points (default = 15)")
return n_points_slider, refresh_interval_slider
@app.cell
def _():
import altair as alt
import marimo as mo
import pandas as pd
import requests
from time import time
from vega_datasets import data
pd.options.display.max_rows = 30
return alt, data, mo, pd, requests, time
if __name__ == "__main__":
app.run()