This post reflects on my first independent Power BI dashboard project and what I learned while building it. Using a Great British Bake Off dataset, I practiced turning raw data into an interactive report while learning how much dashboards depend on cleaning, modeling, design, and revision.
 |
| Great British Bake Off dashboard project. |
In this post:
- Why I Chose This Project
- Learning Power BI While Building
- What I Built
- What Needed Rework
- The Biggest Lesson
- A Few Findings from the Dashboard
- What I’m Studying Next
Why I Chose This Project
After finishing my degree, starting my job search, and continuing to build my technical skills, I wanted to work on a project that felt both useful and enjoyable. Since I am still early in my Power BI learning, I wanted a dataset that would help me practice real dashboard-building skills without losing the fun of exploring the data.
The Great British Bake Off dataset from Tableau’s sample data collection felt like the right fit. It was familiar enough to keep me engaged, especially seeing bakers and seasons I remembered from the show, but it was also complex enough to practice skills that matter in a real dashboard project: loading data, cleaning and transforming tables, building a data model, creating visuals, and designing an interactive report.
This became my first independent Power BI project outside of following course videos and practice activities. Instead of only recreating examples, I had to make my own decisions about what questions the dashboard should answer, how the pages should connect, and how to turn the data into something useful for both Bake Off fans and portfolio reviewers.
Learning Power BI While Building
Before this project, my Power BI experience was limited to one short free online course. That changed quickly when two of my first three interviews asked about Power BI skills. I realized that “I have some exposure to it” was not enough. I wanted to build something more concrete, something I could learn from, document, and eventually include as a portfolio project.
About three weeks ago, I started a Udemy course on Power BI: Microsoft Power BI Desktop for Business Intelligence. I started this dashboard project during the past week as a way to apply what I was learning outside of guided videos and practice exercises. I am currently about 60% through the course. So far, I have worked through loading data, transforming data, and creating interactive visualizations. I am currently in the modeling and DAX section, with Power BI Service still left for the end.
One thing I did not expect was how much dashboard building would bring out my creative side. Power BI and Tableau are technical tools, but they also let me play with layout, spacing, visuals, and flow in a way that feels genuinely fun. It reminded me of being in Jr. high and playing with PowerPoint for the first time, except now I am not just making slides look nice. I am trying to make data easier to explore, understand, and use.
What I Built
The final dashboard has seven pages with several layers of interaction. Users can filter the report by season, gender, age, and placement; drill down into time-based trends; and drill through to more detailed pages for individual bakers or handshake analysis.
The dashboard includes:
Home
The Home page is the landing page for the dashboard. It includes analytical and fan-focused questions that users can answer through the report pages or bookmarked insight views. It also has navigation buttons for quick access to each section.
 |
| Dashboard home page with navigation, reset buttons, and guided insight bookmarks. |
Viewership & Ratings
The Viewership & Ratings page is one of the most analytical pages in the dashboard. It uses line graphs to show trends in the sample rating and sample viewership fields. Users can drill down into the time hierarchy to explore trends by year, month, or day, and use the year slicer to focus on specific time periods.
 |
| Viewership and Ratings page with time-based trends and year filtering. |
 |
| Viewership and Ratings page drilled up to yearly trends. |
Baker Directory
The Baker Directory brings together key information about the bakers. Users can search by baker name and filter by age, gender, placement, or season. From this page, users can right-click a baker and drill through to the Baker Profile page for more detail.
 |
| Baker Directory page with sortable baker details and filters. |
Baker Profile
The Baker Profile page gives a detailed view of an individual baker. It includes their dishes, time allowed for challenges, episode status, outcome counts, and handshakes. This page is designed as a drillthrough destination from baker-focused visuals.
 |
| Baker Profile drillthrough page with detailed baker performance and episode history. |
Demographics
The Demographics page explores relationships between age, gender, placement, and handshakes. The page is fully interactive, with visuals that cross-filter each other when clicked. Users can also drill through from selected demographic views to the Handshake Deep Dive page.
 |
| Demographics page comparing baker age, gender, and handshake patterns. |
Theme Explorer
The Theme Explorer page looks at the themes used throughout the show. It shows how often each theme group appears, what specific themes were included, and the median time allowed for each bake type.
 |
| Theme Explorer page showing episode themes, theme groups, and season filters. |
Handshake Deep Dive
The Handshake Deep Dive is one of my favorite pages in the dashboard. It works as a detailed drillthrough page for exploring Paul Hollywood handshakes by age, gender, placement, season, theme, and baker. This page helps show the relationship between handshakes, themes, and final outcomes.
 |
| Handshake Deep Dive page exploring handshake patterns by season, theme, baker, and outcome. |
Most pages include a reset button that returns the page to its default view, and all pages include a button to return home. I also added bookmarked insight views to answer specific analysis and fan questions more quickly, so users can explore the dashboard on their own or jump directly to guided views.
What Needed Rework
This project did not come together in a perfect order, which is probably exactly what should happen in a first independent dashboard project. I started by getting excited about the visuals, but later realized I needed to step back and strengthen the foundation: the documentation, the data model, and the cleaning process.
One of the biggest changes was the model view. After reviewing the dataset documentation more closely, I updated the model to better match the intended relationships between the tables. That meant some visuals had to be adjusted after the fact, but it also made the dashboard more reliable and easier to explain.
I also changed how I handled missing challenge-time values. Instead of leaving those fields blank, I filled them using the median time for the matching theme group and challenge type. I also added flag columns so I could tell which time values had originally been missing. That felt like a better balance between making the dashboard usable and being transparent about the cleaning process.
The report layout needed revision too. I had an extra page that repeated information found elsewhere, so I reorganized the dashboard and merged repeated ideas into cleaner pages. I also added more useful bookmarks, reset buttons, and analytical questions on the Home page. The rework made the project stronger, and it helped me understand that building a dashboard is not just about making charts. It is about creating a clear workflow from data to model to interaction to insight.
The Biggest Lesson
The biggest lesson I took from this project is that Power BI is much more than building visuals. The visuals are the part people see first, but the quality of the dashboard depends on everything underneath them: the data cleaning, the model, the relationships, the measures, the interactions, and the way the pages are organized.
This project showed me how important it is to slow down at the beginning and make sure the data is ready before building too far ahead. For my next dashboard project, I want to review the documentation more closely first, verify that the data is fully cleaned, check for missing values, confirm that data types are correct, and use clean, consistent naming throughout the model.
I also learned how much the data model affects the entire report. When relationships are not set up correctly, filters, drillthrough pages, and measures can behave in unexpected ways. Going forward, I want to build a cleaner model earlier in the process and make sure the relationships are correct before spending too much time polishing visuals.
The final lesson was about reducing clutter. Some of my early pages repeated information, so I reorganized the dashboard to make each page clearer and more purposeful. This helped me see that a good dashboard is not just about adding more visuals. It is about helping users find answers without making them work too hard.
A Few Findings from the Dashboard
Even though this project was mainly about learning Power BI, the dashboard also surfaced some fun patterns in the data.
The Season 6 finale had the highest sample rating, with a MyRating score of 8.6.
 |
| Highest-rated sample episode bookmark view. |
The Season 7 finale had the highest sample viewership, with approximately 57K sample views.
 |
| Most-viewed sample episode bookmark view. |
DanBH and Rahul received the most Paul Hollywood handshakes, with 3 handshakes each. Both bakers competed in Season 9, which also had the most handshakes of any season.
 |
| Season 9 handshake view showing the season with the most handshakes. |
 |
| Jürgen’s Baker Profile page showing strong non-winner performance. |
By theme group, Biscuits produced the most handshakes.
 |
| Biscuits handshake view showing the theme group with the most handshakes. |
Beyond these individual findings, the dashboard makes it easy to explore details by season, episode, theme, baker, placement, age, gender, and handshake history. Users can start with high-level trends and then drill down into specific bakers or records for more detail.
What I’m Studying Next
Now that this project is complete, I am not finished with Power BI, but I am setting it to the side briefly after spending the last few weeks deep in the dashboard-building process. I still plan to continue improving my Power BI skills, especially with modeling, DAX, and Power BI Service, but I also want to keep building a broader portfolio of completed projects.
One thing I am learning is that I have a lot of project ideas, but I need to be careful not to get pulled in too many directions at once. My next larger project will be a Python hospital readmission project with a capstone-style structure. I plan to follow a full data science workflow: research questions, exploratory data analysis, model building, evaluation, a video presentation, and a final written report.
While working on that project, I will keep learning through Udemy courses, books, blogs, and smaller practice activities. My goal is to keep turning what I study into applied projects that show both technical growth and communication skills.
Thank you for reading my reflection on my first Power BI dashboard. You can dig deeper into the project through the GitHub repository, and I have created a short demo video here to show all the interactions. Comments and feedback are always appreciated.
Comments
Post a Comment