Start with the question, not the chart
Before selecting a type, write the question in one sentence. “Which region sold the most?” is a category comparison. “How did sales change by month?” is a trend question. “What share of support requests came from each channel?” is a part-to-whole question. The question tells you which visual will make comparison easiest.
Use a bar chart to compare categories
A bar chart is usually the safest starting point. It makes differences in size easy to see because every value shares the same baseline. Use it for sales by region, orders by product, tickets by team, or any ranked list. Horizontal bars work well for long labels; vertical columns work well when labels are short.
Sort the pivot table from largest to smallest when rank is the story. Avoid too many series in one chart: if every category has several colored bars, consider filtering the pivot or splitting the view into separate charts.
Use a line chart to show a trend
A line chart is for ordered time periods. Use it when the sequence matters: daily visits, monthly revenue, or quarterly units. Put a valid date field in the pivot rows, group it to a useful interval, and make sure the order runs from earliest to latest. A line connecting unrelated categories such as regions falsely suggests a sequence, so use bars for those instead.
Keep the interval consistent. Comparing one partial month with complete months can create an artificial dip. If there are gaps in the data, decide whether a gap means zero activity or missing information before plotting it.
Use a pie chart only for a small, complete whole
A pie chart can show a simple share of a total, such as a three-part budget or the source mix of a small number of leads. It works only when categories are mutually exclusive and together equal 100 percent. Limit it to a few slices; when readers need to compare similar shares, a sorted bar chart communicates the difference more accurately.
Use the pivot table as the decision check
Before charting, inspect the pivot table. Check that the calculation matches the question: Sum for total revenue, Count for number of records, and Average for a typical value. A well-chosen chart cannot correct an incorrect aggregation or an unclean source table.