— In the News —
As public-health experts have known since the 19th century, information can be the best medicine. What new data streams could help quell future outbreaks?
— Tools and Techniques —
Connected Papers is a visual tool that helps you find and explore academic papers. Enter a title, keywords, a DOI, or paper URL and get back a visual graph that shows connected and related papers that are arranged according to similarity. This looks super useful.
Here are three new database technologies that are worth knowing about. The article summarizes features, things to like, and questions for the project. The related discussion on Hacker News is also worthwhile.
Good introduction to the Scrum development framework and ways it could be applied to data science work.
tslearn is a machine learning toolkit for the analysis of time series data. It's a Python package that builds on scikit-learn, numpy and scipy libraries. The latest release includes a user's guide with a gallery of examples, an API reference, and a quick-start guide.
— Resources —
For a selection of key datasets that are related to the Black Lives Matter protests, police violence, police reforms, police militarization, and tech’s response, check out this issue of the Data is Plural newsletter.
Since January 1, more than 30,000 articles have been published about COVID-19. Here's an overview of some of the best tools being developed that help researchers keep up and find what they need.
— Data Viz —
Is it ever okay to truncate the y-axis of a graph? This post summarizes a recent paper that examines the recurrent debate and why it matters.
Dexplot is a new Python library for creating data visualizations via a simple and consistent API. The design is inspired by Seaborn but it's intended to address some missing features and to be more flexible and easier to use.
— Conferences & Events —
Deep Learning Approaches to Forecasting - This Free Corporate Training Webinar from Metis is an excellent way for managers to learn how deep learning can improve forecasting and planning while avoiding common pitfalls that can easily lead to unreliable results. We will focus on the intuition behind various approaches to modeling to discuss examples, explore techniques as well as how to implement them.