Drill the syntax.
Know it cold.
For data-science interviews, and for the everyday pandas and Python pulls where recalling the syntax is faster than asking AI.
Count rows per category and sort descending. Return a Series.
df.groupby("category").size().sort_values(ascending=False)What you'll drill
The syntax you actually use at work.
df[df["price"] > 100]
df.groupby("user").size()
df["date"].dt.month_name()sorted(items, key=len)
[x for x in xs if x > 0]
{k: v for k, v in d.items()}SELECT region, COUNT(*)
FROM sales
GROUP BY regionnp.where(x > 0, x, 0)
np.unique(values)
arr.mean(axis=0)Prompt
A specific task in plain English.
Answer
Type the one-liner. Hit enter.
Repeat
Wrong drills come back sooner. Right ones space out.
FAQ
Quick answers.
Is it free?
Yes — Syntax Sprint is free during the beta. You need an account so your streaks and spaced-repetition schedule can follow you between sessions.
What syntax is covered?
pandas end to end — filtering, groupby, joins, datetime, string methods, missing data, reshaping — plus NumPy basics and core Python built-ins. New drills are reviewed and added regularly.
How is this different from LeetCode or StrataScratch?
Those test whether you can solve a problem in 20 minutes. Sprint makes the syntax itself automatic: ten-second recall reps, spaced so they stick. Think Anki for pandas — without writing your own cards.
Does my code get executed?
No. Answers are checked as text against accepted forms (quote style, spacing, and common variants are normalized). Nothing you type ever runs.
What if my correct answer is marked wrong?
Most drills accept several equivalent forms. If yours isn't one of them, one click reports the drill and we add the variant — that feedback directly shapes the question set.
How much time does it take?
The daily goal is 10 drills — a couple of minutes. Wrong answers come back sooner, right ones space out over weeks, so short daily sessions compound.