The scikit-learn workbench

The scikit-learn app is a Python editor that runs in WebAssembly (Pyodide). When the status line says scikit-learn is ready, these names already exist:

• np — NumPy
• pd — pandas
• plt — matplotlib.pyplot

You still write from sklearn… imports. sklearn is loaded; the short names are not.

Run code

Paste into the editor and press Ctrl+Enter (Windows/Linux) or Cmd+Enter (macOS). The console is stdout. The Plot panel is for figures.

import sklearn
from sklearn.linear_model import LogisticRegression

print("sklearn", sklearn.__version__)
print(LogisticRegression)

A tiny fit

from sklearn.datasets import make_classification
from sklearn.linear_model import LogisticRegression

X, y = make_classification(n_samples=80, n_features=4, random_state=0)
clf = LogisticRegression(max_iter=200)
clf.fit(X, y)
print("classes", clf.classes_)
print("predict first 5", clf.predict(X[:5]))

A bare clf at the end of the script is not displayed the way Jupyter would. print it.

Files live in /uploads

Click Add files and choose a CSV. After that:

import os

print(os.listdir("/uploads"))

Readable paths: kiosk.csv or /uploads/kiosk.csv.

Writes become download chips

import os
import joblib
from sklearn.linear_model import LogisticRegression
from sklearn.datasets import make_classification

X, y = make_classification(n_samples=40, n_features=4, random_state=0)
clf = LogisticRegression(max_iter=200)
clf.fit(X, y)
joblib.dump(clf, "model.joblib")
print("uploads:", os.listdir("/uploads"))

After Run, click ↓ on the file chip. Nothing is uploaded to Swiftener’s servers.

Privacy

Python runs in this tab. Uploads live in IndexedDB on this device. Lesson pages are public HTML; they never see your models.