By the tiswa editorial team, from the official sources listed at the end of this guide. How we write our guides
You have a table with three student groups and their scores on three tests. You want the chart to show that the second group improved most. You open Excel, select every number, insert a chart, and get ten coloured columns and overlapping lines that nobody can read. The problem is not Excel. You never said which question the chart had to answer before choosing its shape.
Which chart fits your research question?
Begin by deciding what kind of relationship you want to show. Most university projects fall into three cases. You compare separate groups, such as faculties, cities or treatment types. You track one measure over time, such as weekly attendance. Or you show shares of a whole, such as how 300 survey answers split across four options. Each case has a clearer shape than the others, and the wrong shape hides more than it shows.
| Your question | Suitable chart | What to avoid |
|---|---|---|
| How do groups differ? (example: average satisfaction in three clinics) | Column | A pie chart, where small differences look equal |
| How does one measure change over time? (example: attendance across 12 weeks) | Line | Dense columns that break the trend |
| What share does each answer take? (example: 300 survey replies in four options) | Pie or sorted bar | More than five slices in a pie |
| How are two measures related? (example: study hours and grade) | Scatter plot | A connecting line that implies a sequence the data does not have |
The table gives examples to illustrate the logic, not rules for your discipline. Your question decides the choice, and your committee may ask you to justify any shape you use.
How do you build the chart in Excel, step by step?
- Select the data
Choose the cells you want to chart, including the header row if it names your groups. Otherwise the axis shows numbers instead of labels.1
- Open Recommended Charts
On the Insert tab, choose Recommended Charts. Excel suggests chart types based on your data, and clicking any suggestion previews it.1
- Browse every type if needed
If no suggestion fits your question, select All Charts and look through the full list before you decide.1
- Change the type later
Once the chart exists, use the chart options to switch type without rebuilding the data. Keep the source table next to the chart so you can answer questions about any number.1
How do you build the chart in Google Sheets?
The steps are close in Google Sheets. Select the cells, open the Insert menu and choose Chart.2 To change the type afterwards, double-click the chart, open the Setup panel on the right, and pick a new type from the chart type dropdown.2 Whichever tool you use, keep the original table in the same file. When a reader asks where a number came from, the table is the answer.
What mistakes distort the result?
The most common mistake is a vertical axis that starts above zero without anyone noticing. A gap of two points between 78 and 80 looks dramatic when the axis starts at 75. The second is mixing units in one chart, such as percentages next to a count of students. The third is leaving out the sample size, so a clear-looking average hides that it rests on ten people. Check the chart once you have built it with one question: does the shape say what my paragraph says? If not, change the shape or the paragraph.
Is the chart enough on its own in academic writing?
No. The chart shows, and the text explains. Write one sentence above the figure that states the most important number, then one sentence that links it to your question. If the numbers come from an outside source, name that source under the figure. For how to handle the data itself, read the literature review guide. For larger tables, the Excel guide for students covers the basics, and the Excel charts guide covers chart settings inside the program.
Finally, numbering and placement of figures are set by your department, not by Excel. Ask your supervisor whether they require numbered figures and a list of figures at the end of the report.
What do you do when your data comes from a survey?
Say you sent a survey to engineering students at King Saud University in Riyadh and collected 120 replies. This is an illustrative example, not a real result. Before drawing anything, ask how many replies are complete and usable. Write that number under the chart, because it tells the reader how much to trust each bar. If you drop incomplete answers, say so in the text rather than hiding it.
If the question uses a five-point scale, do not stop at the average. An average of 3.2 can hide that half the students chose one answer and the other half chose five. Show the distribution as bars ordered from agree to disagree, and the difference is visible at a glance. That is more honest than one number, and closer to what students actually answered.
Use few colours. Two colours are enough to separate two groups, and every extra shade competes for attention. Many committees print theses in black and white, and a chart that relies on colour alone loses its meaning on paper. Add a pattern such as dashes or dots so the difference survives printing.
Finally, keep a copy of the survey itself. If the committee asks why you chose this question and how it was worded, the chart cannot answer that. The original wording, the number of participants and the dates of collection belong in the methodology section, not under the figure.
