Most business problems I’ve encountered in my career haven’t required predictive analytics or machine learning.
As a Data Journalist –> Data Analyst –> Data Storyteller (or however the market wants to bucket me – I just like to solve data problems and tell compelling data stories), my biggest challenges have been acquiring and cleaning data.
Summary and descriptive statistics, with visual storytelling, have then delivered the insights required. Whether it’s a headline that’s picked up by 100+ media organisations, or providing the data story that aligns commercial teams.
But I’ve always wanted to do something more exciting…
Which is why, I leapt at the chance to use some predictive analytics on a problem. It was part of a case study for a Senior Data Analyst interview.
Specifically, churn analysis – identify which factors are likely to lead to customers churning.
During the next six posts I’ll describe the business problem, and my approach (Association Rule Mining, or Market Basket Analysis). I’ll write about the challenges faced in transforming and encoding the data, and finally the relevance of the insight below.
Stay tuned.
(And if you really want to, you can see my full write up here).
This was first published on my LinkedIn.