— In the News —
Last week in San Francisco, dozens of data scientists from tech companies, governments, and nonprofits gathered to start drafting a data science code of ethics. The general feeling is that it’s about time that the people with powers of statistical analysis woke up to their power, and used it for the greater good. This article by Tom Simonite at Wired explores the effort and why everyone's not on-board.
— Sponsored Link —
The part-time Machine Learning & AI Principles Live Online course at Metis provides an overview of the scientific toolset behind robotics, image recognition, and analytics using a hands-on, project-based curriculum that features popular algorithms and real data sets. The course is taught by popular instructor Sergey Fogelson, VP of Analytics at Viacom. Save $200 on the cost of tuition, only through Feb 20th!
— Tools and Techniques —
Chicisimo offers automated outfit advice, which is complicated in a lot of ways. In this post, founder and CEO, Gabriel Aldamiz-echevarría describes how their data and machine learning approach defines their product and helped them scale from 0 to more than 4 million users to become one of the largest fashion apps in the world. And - they did that in less than 3 years.
In the 2 years since it was initially open-sourced by Google, TensorFlow has rapidly become the framework of choice for both machine learning practitioners and researchers. Last week, J.J. Allaire from RStudio announced a suite of R packages that provide a variety of interfaces to TensorFlow. Here are the details, including links to learning resources and packages.
In his latest post, Denny Britz explores cryptocurrency trading from a Reinforcement Learning perspective. The first half is an overview of how cryptocurrency trading works on the GDAX exchange, data access, and common metrics. The second half describes why Reinforcement Learning should be effective in a trading strategy and how it could work. This isn't a code tutorial but it's an interesting research topic and a worthwhile discussion.
This is a great resource if you ever have an interest in mining, manipulating, and/or visualizing data from Twitter.
FiveThirtyEight is an online news outlet that uses statistical analysis to tell stories about elections, politics, sports, science, economics and lifestyle. Many people don't realize that FiveThirtyEight also shares the data behind each of its articles so you can verify the analysis or dig in and find other stories. This guide to the articles and data is a great learning resource.
Have you already tried these FREE Tools for Developers?
— Data Viz —
Datawrapper recently added a color blindness simulator that automatically flags problematic colors. It's a great idea since most people don't take the time to design visualizations for colorblind users. In this code walk-through, Gregor Aisch shows how it works.
Great guide for getting started with GIS data using R. Begins with a general overview of spatial data and then dives into the two main packages that have standardized the use of spatial data in R: sp and sf. Explores the roots of these packages, how they evolved, the key distinctions, and how to use them.
— Career —
Organizations have realized that success with data requires more than just capable data scientists. You need cross-functional agile teams that include engineers, data visualization experts, and increasingly, "translators." This article in the Harvard Business Review explores how this emerging role fits into organizations and what it takes to succeed.