Tag: power bi

are you team starbucks or team dutch bros?

Either way, coffee isn’t just this magical concoction that many of us love and *ahem* can’t live without — it’s also a dataset waiting to be explored. For my latest project, I built a Coffee Shop Dashboard to dig into customer habits, product sales, and time-based trends.

At first, I planned to keep it simple: load a flat file I found on Kaggle and throw some visuals on a page using Power BI. But the deeper I got, the more I realized this was the perfect chance to practice building a proper data model. What started as a “quick dashboard” turned into a full star schema project. And honestly? That decision made all the difference.

So grab yourself a cuppa joe and as we sip *pinkies out*, I’ll quickly walk you through this project.

my process (with a twist)

1. Data Prep

I used a fictional dataset that came from Kaggle (link here) featuring a coffee shop called Daily Grind Coffee. When I first skimmed the columns, it had everything that I was looking for — a realistic transactional dataset similar to the ones I’ve worked with in the real world. But once I started building from the flat file, new ideas started firing in my head, and I realized I could show even more with the data than I originally planned.

2. From Flat File to Schema

Originally, I was going to connect the flat file directly into Power BI. Instead, I built out a star schema — one fact table (orders) connected to dimension tables (customers, products, dates). Knowing how to build a proper data model is essential for data analysts. A clean schema doesn’t just make dashboards easier to maintain — it makes analysis:

• Scalable: Add new data without breaking your model.
• Flexible: DAX measures are simpler and more powerful.
• Efficient: Queries run faster and avoid messy workarounds.

Having this background is a huge advantage, and it makes me grateful for everything I learned in my database courses in my undergrad. Not every business analyst has formal training in data modeling — many focus on tools and reports without understanding the structure underneath. But knowing the foundation of databases changes the way you think about business questions. You stop just visualizing numbers and start structuring the data so every future question is easier (and faster) to answer.

3. Measures & Metrics

I created DAX measures like average order value, customer retention rate, retained customers, etc.

For example, see the graph below.

This shows the trend of customer retention rate this year. The highest retention rate was in March at 29.91%, which dipped in April and May. Which leads us to asking questions like:

• Did customer retention rate drop due to the warmer climate?
• What items were these customers buying (iced vs hot drinks)?
• What strategies can we take to keep customers buying in April and May? Promotions? Campaigns?

4. design

I kept the dashboard clean and minimal, with aligned visuals, consistent fonts, and filters for exploration.

insights that jumped out

• Croissants and muffins were the surprise stars.
• Customers split into two groups: daily cappuccino loyalists vs weekly tea drinkers.
• Weekends had a completely different sales rhythm than weekdays.

This reminded me that dashboards aren’t just for reporting KPIs — they spark questions you wouldn’t have thought to ask otherwise.

what i learned

• A good schema pays off. That extra effort upfront gave me flexibility and saved time.
• Measures > calculated columns. Cleaner, leaner, and easier to maintain.
• Design is analysis. Layout, spacing, and color choices shape how insights are understood.

💡 CONFESSION TIME:
I’ll admit it — I’m way more comfortable with calculated columns than measures. They feel familiar and straightforward, while DAX has been a tougher learning curve for me. Honestly, I’m still figuring it out.

But that’s part of why I write these reflections. My audience isn’t a room full of experts — to be honest, I’m writing for other analysts like me, all in different parts of their journey in data analytics. And sometimes the most helpful thing isn’t pretending you know everything, but being transparent about where you’re growing.

I’ve learned that while calculated columns get the job done, measures are worth the effort — they make dashboards cleaner, leaner, and much easier to maintain in the long run.

next steps

I’d love to add:

• Customer Segmentation (Daily vs Weekly Buyers).
• Forecasting so the dashboard can move from reporting the past to predicting the future.

Curious about the full project or want the technical nitty-gritty? You can check out the dataset, Power BI file, and documentation on my GitHub repo.

wrapping up: dashboard screenshots

☕ This dashboard gave me a great starting point, but it also got me wondering: what products do customers actually buy together? That’s where market basket analysis comes in. Stay tuned for my next post where I dig into those patterns.

And don’t worry, I won’t end this post without showing you what the dashboards look like!

Voila – a Coffee Sales Dashboard from a Virgo Mind.