— In the News —
Give Google’s DeepStereo algorithm two images of a scene and it will synthesize a third image from a different point of view. That becomes interesting really fast when integrated with Street View. Check out the video. It's amazing.
Working out how to build ethical robots is one of the thorniest challenges in artificial intelligence. Rule-based decisions facilitate implementation but rules are limited by preconceived imaginations. Thought-provoking article.
Google has filed for some machine learning patents recently that are creating a bit of a stir. Is Google being evil or is Google protecting innovation here? This is an interesting Reddit discussion and it's worth being familiar with the issues.
— Sponsored Link —
Check out this new software by Jonathon Morgan and Chris Albon from the Partially Derivative podcast! It's a data discovery tool that enables you to explore your data simply and quickly - no code, no statistics, no spreadsheet wrangling. Lots of examples and sample data makes it easy to dive in. It's currently free and looks awesome!
— Tools and Techniques —
Think you're a good problem solver? This short quiz may surprise you.
Awesome collection of talks from the recent SciPy conference - great speakers and great talks. If you work with Python, it's worth at least skimming this list. You'll likely find something that grabs your interest.
Nice trick. Pandas Categoricals enables you to represent text data numerically to get an order of magnitude speedup on queries. This is a good overview of how to use it.
— Resources —
Text is everywhere and natural language processing is becoming increasingly capable and important. Here's a nice collection of open source tools to help with things like sentiment analysis, topic identification, and automatic labeling of content.
Curated collection of YouTube videos for learning about machine learning, neural nets, and deep learning. There are a lot of interesting picks here and some are just a few minutes long.
Lots of R tools and tutorials here to help gather, analyze, and tell data stories. Contributions are welcome via GitHub. It would be great to see this continue to evolve.
— Data Viz —
Beautiful and hypnotic, these images are adding a new dimension to data analytics. Is it art? And if so, how much value is added to the analytics? Interesting article and be sure to check out the gallery >>
And speaking of artistry, Visual Loop published this interview with Giorgia Lupi this week. Giorgia is a co-founder of Accurat – the hugely popular, data-driven research, design and innovation firm. There are some real gems in this interview, including insights into the future of custom data visualization work.
Collection of 93 data visualization strategies with links to examples and resources. These are organized by strategy types (e.g. correlations, flows, networks, predictive, etc.) and it's easy to find examples that match specific needs. Great resource.
— Conferences —
- The 1st Workshop on Data Systems for Interactive Analysis will be in Chicago in conjunction with IEEE VIS on Oct 26. Submissions due 7/31. www.interactive-analysis.org
- The Big Data Flagship event returns to Boston this September 9 & 10. Hear from the United Nations, Airbnb, Spotify, UPS, eBay & many more. Select the co-located event of your choice and check out the agendas >>
— About —
Data Elixir is curated and maintained by @lonriesberg. If you find this newsletter worthwhile, please help spread the word! Forward to your colleagues or use the links below to share to your favorite network: