DEV Community: Priya Jha

Practice Pandas and NumPy Without Installing Anything (Free Browser Compiler)

Priya Jha — Mon, 28 Sep 2026 03:50:41 +0000

If you want to practice Pandas and NumPy for a data analyst interview, or just get faster at everyday data wrangling, the friction is usually the setup — installing Python, managing environments, getting the right package versions. Here's a way to skip all of that.

I put together a free, browser-based Python compiler (Skillancy Python Compiler) that runs real CPython 3.13.2 (via Pyodide/WebAssembly) directly in your browser tab, with NumPy and Pandas available. No signup, no install, and nothing you write is sent to a server.

Here are eight examples covering the patterns that come up most — from basics to Pandas groupby, merges, missing values, and rolling averages.

1. Basics: variables, f-strings and loops

name = "Skillancy"
scores = [72, 85, 90, 64]
average = sum(scores) / len(scores)
print(f"{name} average score: {average:.1f}")
for s in scores:
    print(s, "pass" if s >= 70 else "fail")

2. List comprehension and counting

collections.Counter is one of the most underused tools in the standard library for quick frequency counts.

from collections import Counter

words = "data analytics needs clean data and clear questions".split()
lengths = [len(w) for w in words]
print(lengths)
print(Counter(words).most_common(2))

3. Functions: dedupe while keeping order

A common interview question — remove duplicates from a list without changing the order of first appearance.

def dedupe(items):
    seen = set()
    result = []
    for item in items:
        if item not in seen:
            seen.add(item)
            result.append(item)
    return result

print(dedupe([3, 1, 3, 2, 1, 4]))

4. Pandas groupby

The bread and butter of any analytics job — summarizing a metric by category.

import pandas as pd

df = pd.DataFrame({
    "city": ["Pune", "Delhi", "Pune", "Chennai", "Delhi"],
    "sales": [120, 90, 150, 60, 110],
})
print(df.groupby("city")["sales"].agg(["sum", "mean"]))

5. Pandas merge

The Pandas equivalent of a SQL LEFT JOIN — combining two tables on a shared key.

import pandas as pd

orders = pd.DataFrame({"order_id": [1, 2, 3], "customer_id": [10, 11, 10], "amount": [250, 400, 150]})
customers = pd.DataFrame({"customer_id": [10, 11, 12], "name": ["Asha", "Ravi", "Meera"]})
merged = orders.merge(customers, on="customer_id", how="left")
print(merged)

6. Missing values: find and fill

Real datasets are messy. This is the standard pattern for spotting and handling gaps.

import numpy as np
import pandas as pd

df = pd.DataFrame({
    "age": [25, np.nan, 31, np.nan, 40],
    "city": ["Pune", "Delhi", None, "Delhi", "Pune"],
})
print(df.isna().sum())
df["age"] = df["age"].fillna(df["age"].median())
df["city"] = df["city"].fillna("Unknown")
print(df)

7. NumPy statistics: z-scores

Vectorized math without writing a single loop.

import numpy as np

data = np.array([12, 15, 11, 18, 20, 14])
print("mean:", data.mean(), "std:", round(data.std(), 2))
print("above mean:", data[data > data.mean()])
z = (data - data.mean()) / data.std()
print(np.round(z, 2))

8. Dates and rolling averages

Smoothing out noisy daily data — a common step before charting a trend.

import pandas as pd

days = pd.date_range("2026-01-01", periods=10, freq="D")
df = pd.DataFrame({
    "date": days,
    "visits": [120, 135, 128, 150, 160, 155, 170, 165, 180, 175],
})
df["rolling_3d"] = df["visits"].rolling(3).mean().round(1)
print(df)

That's the core toolkit for practicing Pandas and NumPy without any local setup. If you want to try these yourself or experiment with variations, the Python compiler is free — it runs CPython 3.13.2 in your browser via WebAssembly, with NumPy and Pandas ready to import.

If there's a Pandas or NumPy pattern you always forget the syntax for, drop it in the comments — happy to add it to the next round of examples.

9 PostgreSQL Queries Every Data Analyst Should Know (Try Them in Your Browser)

Priya Jha — Mon, 28 Sep 2026 03:33:37 +0000

If you're preparing for a data analyst interview, or just want to get sharper at SQL, there's a small set of query patterns that show up again and again: filtering, joins, aggregation, CTEs, and window functions. Get comfortable with these nine and you can handle most of what an interviewer — or a real reporting task — throws at you.

I put together a free, browser-based PostgreSQL playground (Skillancy SQL Compiler) so you can run all of these yourself, no signup or install required. It runs on a small sample database — students, instructors, courses and enrollments — with realistic foreign keys between them, so the joins actually mean something.

Here are nine patterns worth knowing, from basic filtering to recursive CTEs.

1. Filter and sort

The basics: WHERE, ORDER BY, LIMIT.

SELECT name, city, joined_date
FROM students
WHERE city = 'Bengaluru'
ORDER BY joined_date DESC
LIMIT 10;

2. Join and aggregate

Count enrollments per course — a standard one-to-many join with GROUP BY.

SELECT c.title, COUNT(e.enrollment_id) AS enrollments
FROM courses c
JOIN enrollments e ON e.course_id = c.course_id
GROUP BY c.title
ORDER BY enrollments DESC;

3. Anti-join with LEFT JOIN

Find students who haven't enrolled in a specific course. This "LEFT JOIN + IS NULL" pattern is the standard way to find rows in one table with no matching row in another.

SELECT s.name
FROM students s
LEFT JOIN enrollments e
  ON e.student_id = s.student_id
 AND e.course_id = 9
WHERE e.enrollment_id IS NULL;

4. HAVING

WHERE filters rows before grouping; HAVING filters groups after.

SELECT i.name, COUNT(c.course_id) AS course_count
FROM instructors i
JOIN courses c ON c.instructor_id = i.instructor_id
GROUP BY i.name
HAVING COUNT(c.course_id) >= 2;

5. CASE expressions

Bucket continuous values into readable bands — useful for reporting.

SELECT title,
       duration_weeks,
       CASE
         WHEN duration_weeks <= 3 THEN 'Short'
         WHEN duration_weeks <= 4 THEN 'Standard'
         ELSE 'Extended'
       END AS duration_band
FROM courses
ORDER BY duration_weeks, title;

6. Common table expressions (CTEs)

A CTE lets you name an intermediate result and build on it, which keeps multi-step queries readable.

WITH course_counts AS (
  SELECT course_id, COUNT(*) AS enrollments
  FROM enrollments
  GROUP BY course_id
)
SELECT c.title, cc.enrollments
FROM course_counts cc
JOIN courses c ON c.course_id = cc.course_id
ORDER BY cc.enrollments DESC
LIMIT 3;

7. Window functions: RANK

Window functions compute across a set of rows without collapsing them into one row per group — here, ranking each instructor's courses by enrollment count.

SELECT i.name AS instructor,
       c.title,
       COUNT(e.enrollment_id) AS enrollments,
       RANK() OVER (
         PARTITION BY i.instructor_id
         ORDER BY COUNT(e.enrollment_id) DESC
       ) AS rank_in_instructor
FROM instructors i
JOIN courses c ON c.instructor_id = i.instructor_id
LEFT JOIN enrollments e ON e.course_id = c.course_id
GROUP BY i.instructor_id, i.name, c.course_id, c.title
ORDER BY instructor, rank_in_instructor;

8. Window functions: LAG

LAG() looks at the previous row in an ordered set — perfect for month-over-month comparisons.

WITH monthly AS (
  SELECT DATE_TRUNC('month', enrolled_date) AS month,
         COUNT(*) AS enrollments
  FROM enrollments
  GROUP BY 1
)
SELECT month,
       enrollments,
       enrollments - LAG(enrollments) OVER (ORDER BY month) AS change_vs_prev_month
FROM monthly
ORDER BY month;

9. Recursive CTEs

Recursive CTEs generate rows from a starting point — commonly used to build a date series or walk a hierarchy.

WITH RECURSIVE days AS (
  SELECT DATE '2026-01-01' AS day
  UNION ALL
  SELECT day + 1 FROM days WHERE day < DATE '2026-01-10'
)
SELECT day FROM days;

That covers the core patterns that come up most in analytics interviews. If you want to run any of these yourself, or try variations, the SQL playground is free and runs real Postgres (via PGlite/WebAssembly) directly in your browser — nothing to install, nothing sent to a server.

If you've got a favorite SQL interview question that isn't covered here, drop it in the comments — curious what else people get asked.