Customer churn issues are something I’ve tackled in a course before, so it was my first thought of what to do for my portfolio project. Listening to everyone else in the class with their unique ideas– a recommender system, using image data, creating chatbots– helped me look a little more outside of my comfort zone.

If I have something to say for customer churn already, though guided, perhaps I should look deeper into other issues that relate to customers and what a business can do to keep them or improve on their experience.

Perhaps instead of churn, I should look more to what businesses can do to keep the customers that are with them now. What keeps people going to the same bank, is there something that makes them continue to frequent a store in town, or what choice leads them to what gas station over another?

Although I did see a customer churn dataset from Starbucks that included drink orders and thought, what about a recommender system, or a predictive analysis on whether customized drinks lead to churn or not. I think that’s a fun idea. It might be a thought to do that– it’s a small enough idea for me to get through in… Well, seven weeks now.

On the other hand, we were encouraged to look for sets that weren’t already clean, and I have a feeling a dataset from a large company like Starbucks will already be mostly cleaned.

This week’s assignment is to get a dataset, clean it, and get it ready. I’m a little nervous about it. I’m worried about not doing it right or getting the wrong dataset or doing something that’s not going to pan out, but it’s all part of the process, right?

So the process will be to figure out, in better detail, what my questions are, and find datasets for that hypothesis. Then, take the time to clean them. If one thought doesn’t pan out, I can take up the other and will have enough time to clean and poke around on it.

After taking some time to consider this more, I have more or less landed on customer retention. An article I found, here, mentions different types of customer data that can be looked at to help with customer retention. Discovery, Purchase, and Emotional data are the three they mention, and I’m most interested in the emotional data aspect. That probably has something to do with my customer service background.

Okay, that feels narrow enough. Time to find some data to support it.

Like any who read this, I’m hoping my own data cleaning will be swift! I hope you enjoyed a look into my brain on how I’m handling this project week by week.