— In the News —
The storage and transmission of data around the world involves a constant clash of competing local, regional, and national regulations. But why does anyone care where data is physically located?
DJ Patil's latest post introduces an opportunity for the data science community to help define what a Code of Ethics for data sharing should look like. This is a valuable community effort and worth getting involved in.
"Alexa, how is Amazon doing in AI?" The answer? Jeff Bezos’s braying laugh.
— Sponsored Link —
Strata Data Conference in San Jose is where top data scientists, analysts, engineers, and executives converge to shape the future of business and technology. Packed with nearly 200 sessions including executive briefings, industry case studies, and the latest in data engineering, architecture, and machine learning - Strata gives you the tools, skills, and intel to get ahead and drive business results with your data. Save 20% with code DE20
— Tools and Techniques —
Entropy, cross-entropy and KL-divergence are often used in machine learning, in particular for training classifiers. This short video explores where they come from and why they're used in machine learning. This is super clear and well-presented.
Cloud services are the backbone of the modern Internet and AWS is one of the most popular. This tutorial walks through the fundamental AWS services that are useful from a data science perspective. Includes descriptions of the services, specific features that are helpful for working with data, costs, tips and tricks. Even if you already work with AWS, this is a nice reference.
This open-source framework for real-time anomaly detection has an ambitious agenda and is worth watching. The initial release includes support for CSV files, column filtering, a couple of basic modeling and scoring options, and auto-generated visualization.
Oversee your team's entire workflow while organizing your design layouts, coding tasks and assets. Plan visually, make up your timeline and communicate all in one place. Start using the project management tool built specifically for designers and developers.
— Deep Learning —
This reference by Terence Parr and Jeremy Howard aims to explain all the matrix calculus you need in order to understand the training of deep neural networks. No math beyond Calculus 1 is required.
— Data Viz —
Mike Bostock's data visualization projects are consistently very highly regarded. As a graduate student at Stanford, he created the super popular visualization library called D3js and later, he designed visualizations for the New York Times. His latest project is a coding environment for web-based interactive notebooks and it's likely to gain traction quickly.
Strava's global heat map has been big in the news recently because it's been found to reveal sensitive locations. The map itself is beautiful and is definitely worth playing with. For details about how it works, check out this article: The Global Heatmap, Now 6x Hotter