JTBD, Machine Learning & For Loops: Insights & Data Reads

It’s been an incredibly busy few weeks within the Insights & Data team at 383. In November we’ve completed three talks, written two articles and worked on several exciting data projects. Alongside this, we’ve also found time to share some of our most interesting reads (and videos!) from the past few weeks:

JTBD & Milkshakes with Luiza Frederico

Luiza was kind enough to share her subject matter expertise on the Jobs to be Done methodology. You can find Luiza’s slides here and a summary is provided below:

“Products and services are point-in-time solutions for getting jobs done.”

The Jobs to be Done (aka JTBD) framework is all about uncovering customer jobs and identifying met and unmet customer needs in order to inform product and service design, development and innovation.

We strongly believe in the JTBD framework and have applied it across a variety of projects at 383 Project. The following slides give an introduction to the JTBD theory, process and potential applications with key educational takeaways for anyone wishing to learn more.

Agency Applications for Machine Learning and Statistics

This quarter’s HydraHack, a developer event hosted by 383, was all about data science and machine learning, and as such, it made sense to have an Insights and Data team-member there. We talked about agency applications for machine learning and data science, and you can find a video of the talk here. We also published a text version of our talk available here.

Diary of a Data Scientist at Booking.com

We loved this run-through of a typical data scientist’s journey (no holiday pun intended) at Booking.com. Topics covered including distributed computing, regression and classification and that all important part — gadgets to show off your code. You can find the article here.

Why you should forget ‘for-loop’ for data science code and embrace vectorization

Admittedly, this is one really for the data scientists out there, but we thought this comparison of for loops vs vectorization was excellent. Having recently completed a series of Kruskal-Wallis tests which relied on Rvest for some data, we can attest that if you’re write for{i = 0…} in R, you need to rethink your actions.

Like what you read? Give Holly Emblem a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.