Data Analyst Portfolio Project Template: SQL, Metrics, Dashboard, and Case Study

What a portfolio project should prove

A strong data analyst portfolio project should not only show charts. It should prove that you can define a business question, prepare data, calculate metrics, find a pattern, and recommend a next action.

The best portfolio projects feel like analyst work, not homework. "I built a dashboard" is weaker than "I diagnosed which customer segment drove a repurchase-rate decline and recommended a follow-up analysis."

Copy this portfolio project template

Project title:
Business question:
Dataset:
Key tables and fields:
Metric definitions:
SQL workflow:
Dashboard or visuals:
Main finding:
Recommendation:
Caveats:
Next analysis:

Start with a business question

Choose a question that sounds like work an analyst would actually do:

Which channel drove the conversion rate drop?
Which customer cohort has the strongest repeat purchase behavior?
Which product category contributes most to revenue growth?
Where does the signup funnel lose the most users?
Which segment should a retention campaign target first?

Document the dataset

Describe the source, fields, time range, cleaning rules, and limitations. Include a short data dictionary so readers understand what each metric is based on.

Data dictionary example:
orders.order_id: unique order identifier
orders.customer_id: customer identifier
orders.order_date: completed order date
orders.order_value: revenue before refunds
orders.order_status: completed, refunded, canceled
customers.acquisition_channel: first known acquisition source

Show the SQL workflow

Include a query that demonstrates grouping, filtering, joins, and metric logic. The query does not need to be complicated, but it should be connected to the business question.

SELECT
  c.acquisition_channel,
  COUNT(DISTINCT o.customer_id) AS buyers,
  COUNT(DISTINCT o.order_id) AS orders,
  SUM(o.order_value) AS revenue,
  1.0 * COUNT(DISTINCT o.order_id) / NULLIF(COUNT(DISTINCT o.customer_id), 0) AS orders_per_buyer
FROM orders o
JOIN customers c
  ON o.customer_id = c.customer_id
WHERE o.order_status = 'completed'
GROUP BY 1
ORDER BY revenue DESC;

Choose visuals that support the finding

Use a line chart for trends, a bar chart for segment comparison, a funnel chart for step conversion, and a table when exact values matter. Do not include every chart you created.

Write the case study summary

The final project page should read like a concise analyst memo.

Case study summary:
Business question: Which acquisition channel drove the revenue decline?
Method: Compared revenue drivers by channel and decomposed buyers, orders, and AOV.
Finding: Paid social explained 42% of the decline because buyer volume dropped sharply.
Recommendation: Review campaign budget and audience targeting before changing product pages.
Caveat: Attribution is based on first-touch channel and may miss later campaign exposure.

Common portfolio mistakes

Using a public dataset without a business question.
Showing many visuals without a clear finding.
Hiding the SQL and only showing dashboard screenshots.
Ending with observations instead of recommendations.
Ignoring caveats and data limitations.