…We now have over 30 people dedicated to working on analytics, with plans to double that number this year. We’re confident that our systems and processes can scale, and that we can meet any future demand held by the company. In this post, I’m going to share the lessons we learned, and offer a roadmap for other companies looking to scale their analytics function.
This is a stellar read—it’s quite detailed, and this is only the first of four parts (all already published and linked in the intro). If you’re scaling an analytics team today this is your playbook.
People are now building a new kind of software by assembling networks of parameterized functional blocks and by training them from examples using some form of gradient-based optimization. (…) It’s really very much like a regular program, except it’s parameterized, automatically differentiated, and trainable/optimizable.
I apologize in advance for linking to a Facebook post, but Yann does work there so I guess he’s partial.
This was completely new for me—I had never heard of a “spiking neural network”. After reading this post I was interested yet somewhat dubious, but it turns out that there is indeed a niche academic field behind this. The basic concept is more closely mimicking the function of the biological neuron to more effectively sense time series data.
This is not a “use it tomorrow” type of article, but definitely something you should add to your vocabulary and watch in coming months and years.
Data is a representation of real life. It’s an abstraction, and it’s impossible to encapsulate everything in a spreadsheet, which leads to uncertainty in the numbers. (…) Here are some visualization options for the uncertainties in your data, each with its pros, cons, and examples.
From a recent paper at NIPs, and a subject that I found fascinating. Indexes are responsible for the blazing performance in most transactional database systems, but the index data structures they use haven’t changed in many years.
At Fishtown Analytics, we work with venture-funded startups to implement Redshift, Snowflake, Mode Analytics, and Looker. Want advanced analytics without needing to hire an entire data team? Let’s chat.