Ravindra BagaleCourses & study guides

14. Visuals and Their Formatting

14.4 Bar and Column Charts (Stacked, Clustered, 100% Stacked)

Visual Use for Avoid when
Clustered bar (horizontal) Comparing categories, long labels, rankings Time series (use column/line)
Clustered column (vertical) Comparing few categories, short time periods (months) More than ~12 categories with long names
Stacked bar/column Total and composition Comparing inner segments precisely (no common baseline)
100% stacked bar/column Share of composition across categories Absolute size matters

Field wells: Y-axis (category for bars) / X-axis (category for columns), the value axis, Legend (series split), Small multiples, Tooltips.

Examples on our data:

  • Clustered bar: Y-axis = City, X-axis = Total Orders, sorted descending.
  • Clustered column: X-axis = Month Name, Y-axis = Total Sales, Legend = Platform (Blinkit vs Amazon Now side by side).
  • Stacked column: X-axis = City, Y-axis = Total Orders, Legend = Payment Mode.
  • 100% stacked bar: Y-axis = City, X-axis = Total Orders, Legend = Delivery Speed Band. This shows the share of fast deliveries per city.

Key formatting (Visual tab): Y-axis/X-axis (type Categorical/Continuous, range min/max, font, title, Concatenate labels for hierarchies), Gridlines, Bars/Columns (colour per series, Spacing: inner/outer padding), Data labels (position Inside end/Outside end, display units, background), Total labels (for stacked charts), Legend (position, title), Small multiples (grid layout), Zoom slider.

Tip

Mhanje asa: sort bars by value (… › Sort axis › Total Orders › Sort descending), unless the category has a natural order (months, delivery bands).

Practice task

Build a stacked column chart of Total Orders by City with Platform as the legend. Turn on Total labels, then change it to a 100% stacked column and explain what new insight you get.

Ravindra Bagale's Tip

Friends, many students use stacked bars to compare individual segments, but only the bottom segment starts from a common baseline. Use clustered bars when readers must compare segments (UPI vs Card by city), and stacked bars when the total matters most. Sort bars by value unless the axis has a natural order. It's very simple – just make it a habit.