Data Analysis Project Guide: How to Structure and Present Your Work
- Sarah Rajani
- 3 days ago
- 8 min read
A clear format to help you explain your analysis, insights, and impact without confusing your audience.
Table of Contents:
When I started learning data analytics, the part that intimidated me the most was creating projects and writing them up.
We hear it all the time:
Work on projects!
Build a portfolio!
And you might wonder… why?
But if you think about it, it does make sense.
You spent all that time learning the tools, but can you present your information effectively to an audience?
On the job, your stakeholders won't care how you got to your conclusions. The want to know what it means for them.
And that’s where a lot of people get stuck (I did too).
You finish the project, and then you sit down to write about it... and your mind goes blank. Should you walk through every step? Do you explain all your filters and joins? What if you didn’t use a complex tool? Does that make it less valuable?
It can feel like you’re either saying too much or not enough.
The turning point for me was realizing I needed to write effectively and only include what was necessary; not every single detail. That meant changing the way I structured my project writeups, and focusing on what actually mattered to the person reading them.
When hiring managers look at your project, they don’t want to read a technical manual. They want to see how you think.
Once I stopped treating my projects like a checklist and started thinking about it like a story, it became easier. Remember, sharing insights on an analysis means you need to guide the reader through:
A question
A process
A takeaway
Before we get into the sections you should include in your project writeup, let's cover a few quick things first.
Finding Project Datasets
This is the tough part: finding a dataset you want to work with.
There are tons of places to find open source (i.e. free to use) datasets, but if you’re not sure where to start, here are some good places to look:
Kaggle – for structured, beginner-friendly files.
Google Dataset Search – for variety across topics.
Data.gov – for government/public data.
Awesome Public Datasets (GitHub) – lots of categories
Use whatever seems interesting and manageable to you. The best dataset is the one you’ll actually finish a project with.
(I’ll share a full post with more options soon.)
Where to Do Your Writeup
Once you’ve done the project, the next question is: where should you do the write-up?
You have a few options, and they all work. The goal is to make your work visible and readable.
Here are some recommendations:
Posting articles on LinkedIn is one of the simplest ways to do your writeup. You can add a link to the published article in a LinkedIn post and your network can engage and comment. In your writeup, you can link any dashboards (Power BI, Tableau, etc.) you created, and also save your codes in a GitHub README file.
GitHub (in a README.md)
This is a great option if your project is mostly code or SQL-based. Add a short summary at the top of your repo explaining:
What question you explored
What tools you used
Key takeaway
Any links to visuals or dashboards
You can do your full writeup here, and share the code as well. GitHub READMEs are written in markdown, which allows light formatting like bolding and different header sizes.
Even if you choose not to do your writeup here, you should still create a repository of your codes for each project, so that you full codes are available for view.
Medium
If you want to do your writeup on a platform where you can build an audience, Medium is a good place to start. Again, make sure to include a link back to your GitHub or dashboard.
Your Personal Website
If you want to control how everything looks and have one space for all your projects, you can create a personal website. That way you can include:
Screenshots of your dashboard or queries.
A short intro + summary + next steps.
A downloadable PDF or link to a live dashboard if needed.
How to Structure Your Write-Up (Without Overcomplicating It)
This is the part people tend to overthink. There's no hard and fast rule on how to structure your writeups.
But you don’t need a 10-page essay. You just need a logical structure that keeps things focused and skimmable.
Here’s the write-up format I use for my own projects, and what I recommend to anyone building their portfolio:
Title
Keep it simple. Mention the topic, insight, or tool used.
Example:
Tracking Bike Rentals and Weather Patterns in Seattle
Project Context
Before anything else, you need to give some background on the project. In a few sentences or short paragraphs, explain:
Why did you do this project?
Provide some context on the business or dataset your are going to be writing about.
Even if it’s a personal project, make up a realistic use case. Ground it in something someone can follow.
Example:
I looked at subscription data to understand why customers were canceling. The goal was to identify early signals of churn.
Define the Problem or Question
You can include this as part of the Project Context or on its own, depending on how long your introduction is.
But keep this part short and clear. Explain what question you were trying to answer, and why you cared. It can be personal, professional, or curiosity-driven.
What question were you trying to answer?
Who would care about the answer?
Why is it useful?
Example:
This project looks at customer cancellations over a 6-month period to see if there are any early signs of churn. The customer success team needs this information to flag accounts at risk of leaving.
That’s all you need. Just enough for someone to understand what they’re looking at.
Key Findings
Don’t wait until the end to tell people what you found. Most people won’t read that far.
On the job, you don't want to lose your stakeholders' attention in the first few minutes. So in your project writeup, as soon as you’ve explained the context, tell them the key takeaways.
Example:
Users who signed up through referral links were the most likely to cancel: 42% left in the first 30 days. But users who completed onboarding tasks stayed longer. Better onboarding for referred users might help reduce churn.
You don’t need to list everything. Just 1–2 sentences that summarize your insights.
Tools Used
Mention the main tools you used in your analysis and what each one helped you do.
Example:
Excel: quick filtering and data cleaning.
SQL: calculated churn metrics and grouped by segments.
Tableau: built a dashboard showing churn by user type and activity level.
Don't get too technical, just explain your tool selection in a few short sentences.
Walk Through Your Process
This is where most people (me included, at first) start listing everything they did: every column they dropped, every null they fixed, every chart they tried.
You don’t need to include all of that. Just walk people through your main logic. Focus on the steps that actually helped you get to your takeaways.
Try structuring it like this:
What data you used.
Any important filters or definitions you applied.
What comparisons or groups you analyzed.
The key patterns that stood out.
Example:
I filtered the dataset to look at active users in the last 6 months. I created a churn flag for users who hadn’t logged in within 30 days of signing up. Then I grouped users by sign-up method and linked that to onboarding task completion rates to look for patterns.
Keep it short, readable, and focused on the logic, not just the tool.
Use Visuals to Make Your Point
If you’re including charts, they should be there for a reason. Use them to support your takeaway, not to just make your project look good.
A few things that helped me:
Use titles that explain the point, not just describe the chart.
For example: "Early Churn Highest Among Referral Users" is better than writing “Churn by Source”
Highlight the section of the chart you want them to look at.
A great way to do this is by including snippets of sections of your dashboard that you want to highlight, and also provide a link to the full dashboard (Power BI or Tableau Public link).
Make sure to only include the visuals that can add value to your story. Do not paste five charts when one would do the job.
And explain each chart clearly:
What are we looking at?
Why does it matter?
What could someone do with this info?
If your chart doesn’t answer one of those questions, either simplify it, or leave it out.
Main Takeaways
Stick to clear, insight-focused points. This section will mention all the insights you found related to the business question, beyond what you mentioned in number 4 above.
Example:
Referral users churned at nearly double the rate of direct sign-ups.
Completing onboarding tasks correlated with higher retention.
Users with 2+ support tickets were 35% more likely to cancel.
Reflections
Here's where you suggest possible next steps based on your insights. You want to turn your analysis into something useful.
Example:
I’d recommend improving onboarding for referred users and looking into common ticket types that show up before churn.
Limitations
Most datasets and analysis will have some limitations. You need to mention these so that the reader doesn't make assumptions about what the data represents.
No need to over-apologize, just be transparent.
Example:
Didn’t have revenue or pricing data.
Some user IDs were duplicated and removed.
No visibility into mobile vs desktop behavior.
Next Steps
Here you can mention what you’d explore next if you had more time or data. This shows that you're thinking ahead and provides useful recommendations going forward.
Example:
In a future project, I’d want to link this data with support ticket volume. It might help explain whether onboarding issues or technical problems are driving early cancellations.
What You Learned
You wrote up this whole project and did all this work, so you must have learned something. Share it with your readers. Let them know if you had any issues or something took longer than expected.
Example:
I had to learn how to create conditional columns in SQL to define churn. Also learned how much small filters can impact final results.
Invite Feedback or Share a Link (Optional)
This is completely up to you, but some good things to mention might be a short blurb about you and if you're open to comments or suggestions.
Add links to your GitHub, LinkedIn, or portfolio.
Mention where people can reach you.
Keep the tone casual.
Example:
Full code and dashboard are on GitHub. Open to feedback or suggestions!
Final Tips
Your data analysis project write-up is about showing how you think.
Can you take a vague question and break it down?
Can you explain your logic without overwhelming someone?
Can you share an insight in a way that someone else could actually use it?
The goal is to have readers take away something useful from reading your project.
And remember, you don’t need to overexplain your project to make it look impressive.
Use short paragraphs, mention specific takeaways, and write just enough detail to show how you think.
If you structure your write-ups well, you’re not just saying “I know how to use SQL” or “I built a dashboard,” you’re showing that you can think through a problem like an analyst.
That’s what makes your portfolio stand out.
And after you’ve completed a few projects, the next step is to share them. Here’s a post to help you build a great portfolio:
Enjoyed this post? Share your thoughts in the comments.
I write about data, career transitions, and making analytics easier to understand.
Subscribe to my newsletter to get the latest updates.
And come say hi on LinkedIn. I’d love to see what you’re working on.
