Why Every Data Scientist Should Learn Playwright in 2026

# Local page that renders a table with JavaScript, so no external site is used.
import pandas as pd
from playwright.sync_api import sync_playwright

HTML = """
<div id='app'>Loading...</div>
<script>
  setTimeout(() => {
    document.getElementById('app').innerHTML =
      '<table id=t><tr><th>city</th><th>aqi</th></tr>' +
      '<tr><td>Delhi</td><td>180</td></tr><tr><td>Noida</td><td>165</td></tr></table>';
  }, 500);
</script>"""

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page()
    page.set_content(HTML)
    page.locator("#t").wait_for()          # wait for JavaScript to render
    rows = [r.locator("th, td").all_inner_texts() for r in page.locator("#t tr").all()]
    browser.close()

df = pd.DataFrame(rows[1:], columns=rows[0]).astype({"aqi": int})
print(df)   # note: sample values for practice only

Before collecting web data for a project:

[ ] Is there an official API or download? Use it instead.
[ ] Do the terms of use allow automated access?
[ ] Does robots.txt disallow these pages?
[ ] Does the data include personal information? If yes, stop and check law and policy.
[ ] Is the rate slow (pauses between pages)?
[ ] Have I recorded source URLs and collection dates?
[ ] Will I share only what I am allowed to share?

# app.py   pip install flask   Run: python app.py
from flask import Flask, request

app = Flask(__name__)
FORM = '''<form method=post>
<label>Tenure months <input name=tenure></label>
<button>Predict</button></form><p id=result>{}</p>'''

@app.route("/", methods=["GET", "POST"])
def home():
    if request.method == "POST":
        try:
            tenure = float(request.form["tenure"])
        except ValueError:
            return FORM.format("Please enter a number")
        risk = 0.8 if tenure < 6 else 0.2          # stand-in for a model
        return FORM.format(f"Churn risk: {risk:.2f}")
    return FORM.format("")

if __name__ == "__main__":
    app.run(port=5000)

# test_demo.py   pip install pytest-playwright && playwright install chromium
# Start app.py first, then run: pytest -q
import re
from playwright.sync_api import Page, expect

URL = "http://localhost:5000/"

def test_prediction_appears(page: Page):
    page.goto(URL)
    page.get_by_label("Tenure months").fill("3")
    page.get_by_role("button", name="Predict").click()
    expect(page.locator("#result")).to_have_text(re.compile(r"Churn risk: 0\.\d{2}"))

def test_invalid_input_message(page: Page):
    page.goto(URL)
    page.get_by_label("Tenure months").fill("abc")
    page.get_by_role("button", name="Predict").click()
    expect(page.locator("#result")).to_have_text("Please enter a number")

from datetime import date
from pathlib import Path
from playwright.sync_api import sync_playwright

HTML = "<h1>Weekly report</h1><div id='chart' style='width:400px;height:200px;background:#cde'>chart</div>"
out = Path("screenshots")
out.mkdir(exist_ok=True)

with sync_playwright() as p:
    browser = p.chromium.launch()
    page = browser.new_page(viewport={"width": 1280, "height": 800})
    page.set_content(HTML)   # or page.goto(your_dashboard_url) when authorised
    page.locator("#chart").screenshot(path=out / f"chart_{date.today()}.png")
    page.screenshot(path=out / f"report_{date.today()}.png", full_page=True)
    browser.close()
print(sorted(p.name for p in out.iterdir()))

from playwright.sync_api import sync_playwright

HTML = "<h1>Model results</h1><p>Churn risk by plan</p><table><tr><td>Basic</td><td>0.42</td></tr></table>"

with sync_playwright() as p:
    browser = p.chromium.launch()
    phone = p.devices["iPhone 13"]
    context = browser.new_context(**phone)
    page = context.new_page()
    page.set_content(HTML)
    page.screenshot(path="mobile_view.png", full_page=True)
    browser.close()
print("saved mobile_view.png")

Project: Tested model demo with automated reporting

- Flask or Streamlit demo app serving a simple model
- Playwright end-to-end tests: valid input, invalid input, page loads
- GitHub Actions running the tests on every push
- Script capturing weekly screenshots of results pages
- Optional: permitted data collection from a local or open-data page
- README covering responsible data use and how to run tests

Week 1  Playwright basics: goto, locators, fill, click, expect
Week 2  End-to-end tests for your model demo; run in CI
Week 3  Screenshots, optional permitted data collection, README

Situation: a model demo occasionally broke before meetings
Task: make the demo reliable (practice project)
Action:
  - Playwright tests for the full prediction flow
  - tests run automatically in CI
  - screenshots captured for weekly reports
Result: demo checked automatically; I can explain each test