Portfolio · Week 1 of 4 · 5 hours
Portfolio Planning: Dataset Selection, Problem Definition and Data Collection
Design a portfolio that shows a range of skills, and start the new projects.
Learning objectives
By the end of this week you will be able to:
- Explain what employers look for in a data analyst portfolio.
- Plan three projects that cover different skills and industries.
- Choose datasets and write a problem definition for each.
- Collect the data and set up a consistent folder for each project.
Lessons
What Makes a Strong Portfolio
A portfolio is evidence. Hiring managers usually spend only a few minutes on it, so it must show quickly that you can take messy data, ask good questions and communicate answers. Quality beats quantity: three solid, well-documented projects are better than ten half-finished ones.
Plan for range. Together, your projects should show:
| Project A | Project B | Project C | |
|---|---|---|---|
| Main tool | Excel dashboard | SQL + Power BI | Python (pandas) EDA |
| Industry | Retail sales | Healthcare or HR | Your target industry |
| Question type | Performance (what happened?) | Comparison (who or what differs?) | Drivers (what is associated with the outcome?) |
Your Project course capstone can fill one column. If you are targeting a specific industry, make at least one project about it.
Dataset Selection
Use the same checklist as the Project course (size, documentation, license, privacy) and add two portfolio-specific tests:
- Is it fresh? Avoid the datasets every bootcamp uses (Titanic, Iris, the standard Superstore file) unless you take an unusual angle.
- Is it relevant? Data from the industry you want to work in shows interest and domain knowledge.
Good sources: Kaggle, government open data portals, city open data, Our World in Data, the World Bank, and public company or sports data.
Problem Definition
Every portfolio project should open with a clear business problem, because that is what separates an analysis from a tutorial. For each project write:
- A one-sentence problem statement naming a stakeholder (it can be a realistic imagined one: "the operations manager of a city hospital").
- Three SMART questions.
- The decision your analysis supports.
Readers will judge the whole project by these first lines, so make them specific.
Data Collection
Use one consistent folder structure for every project so your portfolio looks professional: data/raw, data/clean, sql, notebooks, dashboard, images, plus README.md. Record sources and licenses as you go.
A note on file size: GitHub limits single files to 100 MB. For large datasets, include a small sample in the repository and link to the full source in the README.
Watch
- VideoCOMPLETE Data Analytics Portfolio Project in 6 EASY Steps
Overview of planning and building a portfolio project.
Practice
Fill in the portfolio planning table for your three projects. For the two new projects, choose datasets, write problem definitions and SMART questions, and set up the folders with the raw data.
Check your understanding
Try each question first, then open it to see the answer.
Why are three strong projects better than ten weak ones?
Reviewers spend little time, and depth, documentation and quality show job readiness better than volume.
Name two ways to give a portfolio range.
Any two of: different tools, different industries, different question types.
What should you do if a dataset is larger than 100 MB?
Commit a sample and link to the full source in the README.
Assignment
Submit your completed portfolio plan (table, problem definitions and dataset sources) for instructor feedback.