Descriptive Statistics
Senior 4 · Basic Mathematics
Descriptive Statistics
Lesson 2 of 5
25 min
The 2022 census gives each province's population: City of Kigali 1 745 555, Southern 3 002 699, Western 2 896 484, Northern 2 038 511 and Eastern 3 563 145 (NISR, Table 1). A planner may want to compare the provinces with each other, or to show each province's share of the whole country. A geographer may want the shape of the 30 district populations from the last lesson. Each question needs a different graph, and choosing the wrong one can hide the answer or even mislead.
Definition
Bar chart: separate bars of equal width whose heights show the frequencies of categories or discrete values; gaps between bars. Pie chart: a circle divided into sectors whose angles are proportional to the frequencies:
Histogram: touching bars over the class boundaries of grouped continuous (or treated-as-continuous) data, whose areas are proportional to the frequencies. Frequency density . Frequency polygon: the midpoints of the tops of the histogram bars joined by straight lines.
The bar chart makes comparisons easy: the Eastern Province is the most populous, the City of Kigali the least, and the Southern and Western Provinces are almost equal. A pie chart shows the same data as shares of the whole.
For grouped numerical data, the bars of a histogram touch, because the classes run on without gaps. When all classes have the same width, bar heights can show the frequencies directly.
The shape is skewed to the right (positively skewed): most districts are bunched at the low end, and a long thin tail stretches to the right, ending with Gasabo. A frequency polygon shows the same shape as a line, and is useful for comparing two distributions on one grid. It is closed by joining it to the horizontal axis at the midpoints of the empty classes on each side (250 and 950).

Tip
Choose by the question. Compare categories: bar chart. Show parts of a whole: pie chart (five or six sectors at most). Show the shape of grouped numerical data: histogram. Compare two distributions: frequency polygons on one grid. And always give a title, labelled axes with units, and the source.
Worked example 1: sector angles
Calculate the sector angle for each province for a pie chart of the 2022 population (total 13 246 394).
Similarly Southern , Western and Northern . Check: ; the missing is rounding, so the angles are consistent with .
Worked example 2: unequal class widths
A report groups the 30 districts as (15 districts), (10) and (5), in thousands. Draw a correct histogram.
The last class is four times as wide. Measure widths in units of 100 thousand and divide:
Why it matters: a bar of height 5 over the wide class would have an area of units, larger than the 400–500 bar (10 units), suggesting more districts above 500 000 than between 400 000 and 500 000, which is false. With frequency density, area : correct. The frequency of any class can be read back as density × width.
Worked example 3: interpreting a graph
Use the histogram of district populations and the density map to write three statements a planner could use.
Each statement quotes a figure from the graph and says what it means: that is what "interpret" requires.
Warning
Misleading graphs. A bar chart whose vertical axis starts at 1.5 million instead of 0 makes Kigali look tiny next to the other provinces. A histogram with unequal classes drawn with frequency, not density, distorts the shape. A pie chart with too many sectors, or drawn in 3D, makes angles hard to compare. When you read a graph, check the axis scales, the class widths and the source before you believe it.
Corbettmaths, "Reading Histograms". Watch how frequency = frequency density × class width, then use it to read back the 5 districts in the wide class of worked example 2.
Practice
Try each one before opening the answers. They get harder as you go.
Key points
Sources and credits
Quick check