Chapter 8 · Data Analyst
Visualization and dashboards
~5 min read
A chart's job is not to be pretty. It is to make the right comparison obvious in seconds. Most bad charts come from picking the wrong type for the message. This decision guide covers the large majority of cases you will meet.
8.1 Chart selection#
| You want to show | Use | Avoid |
|---|---|---|
| Comparison across categories | Bar or column | Pie with many slices |
| Change over time | Line | Three-dimensional anything |
| Relationship between two variables | Scatter | Dual-axis tricks |
| Part-to-whole with few parts | Stacked bar or a single pie | Pie with eight or more slices |
| Distribution | Histogram or box plot | The average alone, which hides spread |
8.2 Principles behind good dashboards#
- Lead with the answer. The most important number goes top-left, where eyes land first.
- One dashboard, one question. "How is the business doing" and "why did signups drop" are different dashboards.
- Reduce ink. Remove gridlines, borders, and decoration that carry no information. Less chrome, more signal.
- Always give numbers context. A metric alone means little. Show the comparison: versus last period, versus target, versus segment.
Before you ship any chart, ask what the viewer should do with it. If you cannot answer, the chart is not ready. Tools like Tableau and Power BI are worth learning, but the tool is never the point. We would rather hire someone with a clear story and a plain bar chart than someone with a beautiful dashboard and no point of view.
Get the next chapter and weekly interview tips by email
One short email per week. Skim in a minute. Unsubscribe anytime.