The Task
The task for day one of dashboard week, was to web scrape data from the Playbill Grosses List. This involved extracting weekly Broadway show performance data and using it to uncover insights through creating a Tableau dashboard.
Web Scraping
Before I began web scraping, I wanted to have a general idea of what I wanted to create the dashboard from first. I came to the conclusion that I wanted to focus on one show, The Book of Mormon, and get all the relevant data to the show. I didn’t have too many issues with the web scraping aspect, only that I struggled to get the regex working to download all of the relevant pages. Due to the time constraint I decided to hard code it in using the generate rows tool. This managed to achieve the desired results but led to a workflow that is no longer dynamic. If I want to retrieve updated data in the future, I’ll need to manually adjust the conditions within the workflow.
Cleaning the Data
Cleaning the data at first, I thought was rather straight forward. I ran it through all the usual tools, thought everything looked clean and went straight into Tableau for dashboarding. However, once I started developing the dashboard, I quickly noticed that the weekly gross figures were far too high, and some play titles contained unexpected HTML tags. After looking into the source, I found that the websites table structure had two different records combined into a single row. The recent entries had missing values entirely, which caused me to completely overlook the issue during the initial clean. Fortunately, it didn’t take me long to clean but reminded me that I need to watch out for multi row entries in the future.
Dashboarding
For the dashboard I wanted to look at The Book of Mormon and notably, how it has been performing over time since its opening. The goal was to see whether the show has maintained the strong momentum it built from its debut in 2012, or has the shows demand gradually declined over the years.
Writing the calculations for the dashboard turned out to be quite time consuming. I found myself struggling at times, especially as I kept coming up with new ideas and tinkering with different approaches. While it did add to the workload, it helped me uncover deeper insights. The insights did find evidence that there has been a decline in nearly all metrics over the years such as capacity percent, weekly gross, average seats sold and average ticket price.
Although all that time spent on building the calculations behind the dashboard, left me with very limited time for formatting and designing. This led to a final dashboard that I wasn’t too happy with, at least in terms of formatting. For the coming days, I’ll need to work on balancing the time spent on data logic with the time needed for visual design to ensure that future dashboards are visually polished and engaging.
In Summary
I am quite happy with what I created, both in terms of the insights found and functionality of the dashboard itself. That said, I definitely plan to revisit this project once dashboard week has concluded to enhance the visual appeal and bring more life to the overall design.
