Google Data Analytics

June 2023

I completed the Google Data Analytics course as my first serious foray into the world of data.

The course is fairly comprehensive, though that means it cannot delve too deep into any singular aspect of data analysis. It covered the process and ethics of gathering data, asking the right questions, choosing the right tools and metrics, and bringing it all together to deliver meaningful conclusions.

The course gives some training in specific tools:

• Excel and Google Sheets (I am already quite proficient)

• Tableau and POWER BI (Fun visualization tools I intend to explore further)

• R (programing language and libraries useful for data visualization)

Capstone Project

I chose to go with a provided case study and dataset for the capstone project which ties it all together. The dataset has ride information from the past year for a bikeshare company using real data scrubbed for any personal or identifying information.

The main question being asked is how bikeshare service subscribers and casual users (no-members) differ in their use of the service. This is in an effort to convert casual users into members.

I decided to go with POWER BI as my tool for cleaning and visualizing the data. I am already familiar with Power Query from Excel, and POWER BI uses the same tool to gather and clean the data. However, it offers a much more powerful engine for slicing and visualizing the data. It was fun to dig around and find out what my options were with this unfamiliar tool.

I found this dataset was actually a bit too flat to really take advantage any advanced tools. There is only one major factor by which the data is being sliced: membership status. Every metric can be broken into its own table or visualization sliced by this one factor, so drilling down by selecting an attribute to slice by doesn't really impact the other charts significantly. Perhaps that is simply a quirk of this data in that only one significant difference was found: members ride more during weekdays and casuals ride more during the weekends. Every other metric was virtually identical between the two groups, aside from there simply being more rides by members in general.

That is the answer to the main question: how do they use the service differently. The course recommends coming up with ideas for how this information can be used to convert casuals into members, but, personally, I don't think that is the role of an outside analyst. Now if I worked for the rideshare company I might brainstorm some ideas, such as making sure the bikes are available near all local business centers and near recreational areas, and provide bikes that can allow for transportation of goods or groceries. But if I was hired as an outside analyst, I would prefer to stick to the pure observations that can be made from the data, as I am not an expert on bikeshare businesses.

The Visualizations

Now, despite my slight objection to what the Google course was asking with this project, I still greatly enjoyed coming up with new ways to display the correlations I had found in the data.

I was particularly happy with this pair of doughnut charts displaying the disparity between member and casual riders during the weekdays vs during the weekends connected by a bike frame.

A bit more toward an interactive infographic instead of straightforward visualization, but I thought it was fun. Why not have visualizations be interesting and evocative if you have the time to pretty them up a bit.