Our mission is to help data scientists deliver value for their organizations.
What we do
The vast majority of data science training, whether it be traditional courses, MOOCs, bootcamps, or one-off blog posts, is about methods. Sometimes the methods are simple, sometimes they're wildly complex, but it's almost always about the methods.
Then, data scientists go to work for organizations that ask them to solve business problems. Things like
- How to best allocate the marketing budget?
- How to roll out a new product feature in a way that doesn't jeopardize customer satisfaction?
- How to prevent our customers from churning?
- What are the top bits of feedback in our app store reviews?
- Etc, etc
Our goal is to bridge this gap by writing about data science from a perspective that keeps the business problem foremost in mind. We hope this curated, hands-on, deep-dive content will help you to level up your own data science practice.
What we don't do
We do not accept referral or affiliate fees for reviews, which means our content is based entirely on our experiences and opinions as practicing data scientists. This philosophy not only avoids direct biases, it also means we are very open to feedback from our readers; if you see an angle we've missed or screwed up, please let us know!
Who we are
Brian Kent, Founder
Since receiving a Ph.D. in statistics from Carnegie Mellon, I have worked as a practicing data scientist in a variety of roles and companies.
- At Turi (née GraphLab), I built high-performance machine learning toolkits for other data scientists to use.
- At Apple, I functioned as an internal machine learning consultant, advising and prototyping for several mission-critical areas in silicon engineering, health platform, and autonomous systems.
- At Credit Sesame, I served first as machine learning team lead, then as Director of the Data Science and Analytics group.
Part of my progression has been a journey from "tool-builder" to "internal tool-user" to "customer-facing tool-purchaser"; from each of these perspectives, I have seen the gap between data science methodology and real-world business needs. In some cases, the methodologists fail to understand business objectives, customs, or mental models; in others, the business operators do not understand or trust data science methods to help them reach their goals.
My focus now is closing this gap to enable the best possible data-guided decision-making in real-world applications. I'm very excited to share my next explorations with you in the form of The Crosstab Kite.
What does Crosstab Kite mean?
A Crosstab is a contingency table, one of the most foundational methods for making decisions with data. Crosstabs are fast and simple, which is usually more important for applied problems than tinkering with state-of-the-art research methods.
As a field guide to data science, we take inspiration from more traditional field guides, like the Audubon Guide to North American Birds. The swallow-tailed kite is our favorite entry, and our mascot and logo.