An analyst portfolio should show how you turn a question into a supportable explanation. The chart is only one part of that work. A reader needs to understand the data, the checks and what the analysis cannot establish.
Use public data with appropriate permissions or clearly labelled synthetic data. Do not publish a former employer’s private records.
1. A service-volume summary
Use synthetic support records to describe volume by period or category. Explain the definitions and missing information before drawing conclusions about workload.
2. A data-quality review
Identify duplicates, inconsistent labels and missing fields in a practice dataset. Show how decisions affect the result rather than silently removing inconvenient records.
3. A simple operational dashboard
Choose a few measures linked to a question. Define the denominator and period. Avoid displaying precise-looking percentages when the sample is too small to support a strong interpretation.
4. A before-and-after comparison
Explain what changed and what else might affect the outcome. A difference over time does not by itself prove that one action caused it.
5. A public-dataset explainer
Summarise an official dataset’s scope, units and reference period. Keep a link to the source and avoid presenting an old survey as a live count.
6. A reproducible analysis notebook
Record the steps clearly enough that a reviewer can follow them. Keep secrets and private paths out, and state any environment assumptions without overcomplicating the project.
7. A decision memo
Write a short explanation of the finding, uncertainty and next question. Show how you would communicate to a colleague who does not need every technical detail.
Choose the next practical step
Choose evidence relevant to the role rather than including every chart you have made. Be ready to explain a cleaning decision and a limitation. The analyst-versus-scientist guide helps you compare role scope, while the achievement guide helps describe your own contribution accurately.
