Wheel of Churn
The goal of this project was to build an interesting data visualization displaying customer churn that showed the capabilities of the data science and web development team. The data viz would take the output of a predictive analytics software and display it ina data viz. This was to be presented to a large audience at an upcoming conference and also attract new clients.
I was responsible for the design of visualization and worked with a developer and data scientist who built the design.
Understanding the data
I met with our data scientist and web developer to understand vision for this project. They were using customer data that predictive analytics software would use to predict the likelyhood of a customer churning. Customer churn is the customers who would stop using the company's services. I had to take this output from Emcien and figure out an interesting way to visualize it. The image below shows an example of the output.
We wanted something exciting that would spark conversation and help the teams network with more clients. We had an initial brainstorming session together, the outcome of which was a node diagram that showed different attributes. I took this initial idea and pushed it further.
Each line on here represents a rule. An example of a rule is the contract length is 2 years long. A line that has more red than blue in it indicates the customer has a higher probability of churning.
If there are two circles at the end of the line then it represents a combination of 2 rules. This would indicate if a customer is in their first year of contract and the contract length is 2 years long, they have a 90% chance of churning (for a line that was mostly red).
Memorability vs Readability
One of the takeaways from this project was that designing something unique also requires a lot more effort from the readers part. It can be confusing to people or take long for them to understand it but it's also more memorable. Sometimes this trade off might be worth it depending on the purple of the visualization.