Step 01Find the overlap
If you want a data project that shows what you can do, start with three job descriptions for roles you actually want. Mark the skills, tools and business tasks that keep coming up, then find where they overlap. That overlap tells you which direction to take, without copying any one posting.
Step 02Turn it into one question
Next, write the direction as a question you can answer, find data that fits it, and limit the scope. For example: organise a sales dataset and make one recommendation. Before you publish a conclusion, check the dates the source table covers, so a change in the numbers is not just a missing day of data. Choose tools that serve the question, not tools that fill out a checklist.
Step 03Tell it as a story
Finally, tell the result as a story someone can follow quickly: what the question was, what the data could not tell you, which judgements you made and what the result means. The three job descriptions give you a focused start; the project itself still rests on your question, your evidence and your reasoning.