Pandas Top 20 Functions for Data Analyst Interviews 2026 | Data Analyst Interview
🐍 Pandas Reference

Pandas — Top 20 Functions for Data Analyst Interviews 2026

The 20 Pandas methods that appear in 80% of data analyst interviews — explained with real code examples you can run and adapt immediately.

Book Free Session ← Python Hub
20 Functions
Master these and you handle any Pandas interview question

Every function below has appeared in real interview questions at top companies in 2025–2026.

Easy

1. df.head(n) / df.tail(n) — preview first or last n rows

Exploration · Always the first step when receiving a new dataset
Easy

2. df.info() — column names, dtypes, non-null counts

Exploration · Tells you which columns have nulls and wrong types in one call
Easy

3. df.describe() — summary statistics for all numeric columns

Statistics · Count, mean, std, min, 25th/50th/75th percentile, max
Easy

4. df.value_counts() — frequency count of a categorical column

Exploration · df[‘category’].value_counts() — add normalize=True for %
Easy

5. df.isna().sum() — count missing values per column

Data Quality · df.isna().sum() / len(df) * 100 gives % missing per column
Medium

6. df.loc[rows, cols] — label-based selection

Selection · df.loc[df[‘city’]==’Mumbai’, [‘name’,’revenue’]] — label + condition
Medium

7. df.groupby().agg() — group and aggregate with named results

Aggregation · df.groupby(‘region’).agg(revenue=(‘amount’,’sum’), orders=(‘id’,’count’))
Medium

8. pd.merge(df1, df2, how, on) — join two DataFrames

Joining · how=’inner’/’left’/’right’/’outer’ — mirrors SQL JOIN types exactly
Medium

9. df.sort_values(col, ascending=False) — sort rows

Sorting · .head(10) after sort_values gives top-N — extremely common interview pattern
Medium

10. df.fillna(value) / df.dropna() — handle missing values

Cleaning · fillna(df[‘col’].median()) for numerical, dropna(subset=[‘key_col’]) for rows
Medium

11. df.drop_duplicates(subset) — remove duplicate rows

Cleaning · subset=[‘user_id’,’date’] keeps first occurrence of each user per day
Medium

12. df[‘col’].str.contains() / str.replace() — string operations

Strings · Vectorised string methods — no loop needed. df[df[’email’].str.contains(‘@gmail’)]
Medium

13. pd.to_datetime() + dt accessor — date parsing and extraction

Dates · df[‘date’].dt.year, dt.month, dt.day_name(), dt.to_period(‘M’)
Medium

14. df.pivot_table(index, columns, values, aggfunc) — reshape wide

Reshaping · Monthly revenue by category in one line — staple of take-home tasks
Medium

15. df.rename(columns={}) / df.assign() — rename and add columns

Wrangling · df.assign(margin=df[‘revenue’]-df[‘cost’]) — cleaner than df[‘col’]=
Hard

16. df.groupby().transform() — broadcast group aggregates back to rows

Advanced · df[‘pct_of_total’] = df[‘revenue’] / df.groupby(‘region’)[‘revenue’].transform(‘sum’)
Hard

17. df.apply(func, axis) — apply custom function row or column wise

Advanced · Use only when no vectorised alternative exists — show interviewer you know the performance trade-off
Hard

18. pd.cut() / pd.qcut() — bin continuous variable into categories

Segmentation · pd.cut(df[‘age’], bins=[0,25,40,100], labels=[‘Young’,’Mid’,’Senior’])
Hard

19. df.melt(id_vars, value_vars) — unpivot wide to long format

Reshaping · Opposite of pivot_table — turns column headers into rows for tidy data
Hard

20. df.shift() / df.pct_change() — lag and growth rate on time series

Time Series · df[‘mom_growth’] = df[‘revenue’].pct_change() * 100 — one-line MoM in Python

Want to practise these in a live Pandas mock?

Book a free Python mock — we test you on groupby, merge, pivot, and time series questions just like real interviews.

Book Free Python Mock
Subscribe
Subscribe