Missing values

Real data is rarely complete. A missing value shows up as NaN ("Not a Number") -- written as None when you build a DataFrame by hand, pandas converts it automatically:

data = {"x": [10, None, 20, None, 5]}
df = pd.DataFrame(data)
print(df)

Finding missing values

print(df["x"].isna())        # True/False per row
print(df["x"].isna().sum())  # how many are missing, as a count

Filling missing values

filled = df["x"].fillna(0)
print(filled.sum())

.fillna(value) replaces every missing entry in that column with value -- it doesn't change the original column unless you assign the result back to it (df["x"] = df["x"].fillna(0)).

Creating a new column

Any expression using existing columns can be assigned straight to a new column name:

df["score_x2"] = df["score"] * 2

This runs on every row at once -- no loop needed.

Why This Matters for Your Exams

Ignoring missing values instead of handling them deliberately is the most common real-world data analysis mistake -- a .sum() or .mean() over a column with NaNs silently skips them rather than erroring, which can quietly change your answer if you weren't expecting gaps in the data. Always check .isna().sum() before trusting a summary.

Try it yourself