Comparative Bar Charts and Time Series Data

Comparing groups with bar charts

A bar chart shows a value, often a frequency, for each category. If the same categories are recorded for two groups, a multiple bar chart places their bars side by side. For example, a survey might record how pupils in Year 7 and Year 8 travel to school. Beside the label bus, one bar could show 18 Year 7 pupils and the other 20 Year 8 pupils. A key identifies the groups, and the common vertical scale lets you compare their values directly.

A composite bar chart, sometimes called a stacked bar chart, shows the same information differently. Each category has one bar divided into labelled sections. The whole bar shows the category total; each section shows one group's contribution. If 18 Year 7 pupils and 20 Year 8 pupils travel by bus, the bus bar reaches 38. The Year 8 section extends from 18 to 38, so its value is $38-18=20$, not 38. Composite bars make totals and their make-up easy to see, while side-by-side bars make direct group comparisons easier.

Classifying information in a table

Before drawing a chart, organise the observations into a table. Decide what each classification means so that every observation has a clear place. For the travel survey, put walk, bus and cycle in rows and Year 7 and Year 8 in columns. Count pupils in each row–column combination; a pupil who cycles in Year 7 belongs in the cycle, Year 7 cell. Row totals show how many pupils chose each way of travelling, and column totals show the size of each year-group sample. Those counts can then be used for either comparative bar chart.

Showing change over time

When observations are made at successive times, their order matters. A time-series line graph puts time on the horizontal axis and the measured quantity on the vertical axis. Choose even scales, label the axes and units, plot each observation at its time, then join successive points with straight lines. Unlike separate category bars, the joined points help your eye follow how the quantity changes.

Data for Illustrative quarterly profit over two years

Read a time series at two levels. First, look for its overall trend: is the quantity generally increasing, decreasing or staying roughly steady? Then look for shorter-term patterns. In the graph, profits are higher in each quarter of year 2 than in the matching quarter of year 1, so the overall trend is upward. The rise to quarter 2, dip in quarter 3 and rise again to quarter 4 occurs in both years. That repeating pattern is seasonal variation. A line between quarterly observations helps display the pattern; it does not mean a value was actually recorded between those quarters.

Comparative bar charts

Multiple: bars for groups sit side by side within each category; compare their heights using the key. Composite: one bar is split into sections; its full height is the total, and a section's value is its upper boundary minus its lower boundary.

Tables and time series

Classify each observation in the appropriate table cell and use row or column totals to check counts. For a time-series graph, put time on the horizontal axis, plot observations in order and join successive points. Describe the overall trend and any regularly repeating seasonal variation separately.