This lesson's interactive features are locked, please to keep using them

Pivot Tables

A pivot table reshapes data from long format (many rows, few columns) to wide format (fewer rows, more columns). Think of it as a groupby that spreads one column's values into new columns.

Pivot tables are often exactly what you want in a report or dashboard – they map 1:1 to what stakeholders expect to see.

Long vs. Wide Format

In long format, every observation is a row:

team	priority	response_minutes
Support	high	45
Support	low	12
Billing	high	60
Billing	low	20

In wide format, categories become columns:

team	high	low
Support	45	12
Billing	60	20

Creating a Pivot Table

Use pd.pivot_table():

tickets = pd.DataFrame(
    {
        "team": ["Support", "Support", "Billing", "Billing", "Support", "Billing"],
        "priority": ["high", "low", "high", "low", "high", "low"],
        "response_minutes": [45, 12, 60, 20, 30, 25],
    }
)

pivot = pd.pivot_table(
    tickets,
    values="response_minutes",
    index="team",
    columns="priority",
    aggfunc="mean",
)

Pivot for presentation, but keep long format for computation. Long format is usually easier to work with for further analysis.

Handling Missing Combinations

If some group combinations don't exist, you'll get NaN. Use fill_value to replace missing combinations:

pivot = pd.pivot_table(
    tickets,
    values="response_minutes",
    index="team",
    columns="priority",
    aggfunc="mean",
    fill_value=0,
)

Assignment

SnackStack's operations team wants an energy dashboard: total power draw for every region, broken out by device type.

Complete the energy_grid_by_region_and_type function. It accepts a df of telemetry rows with region, device_type, and energy_kwh columns, and returns a pivot table.

Pivot the data with region as the index and device_type as the columns.
Total the energy_kwh values for each cell by summing them.
Fill any region/device-type combination that has no readings with 0.
Return the grid with region as the index (don't reset it).