There’s nothing tastier than a set of links to devour! Check out what we’ve been reading recently and be sure to share your links to thought-provoking articles and discussions on topics in the field of data science.
Vice News’ “Parallel Narratives” is a really interesting analysis of how the 2016 US election was extremely polarized on social media. It argues that mainstream media didn’t foresee Trump winning because journalists’ information bubbles were devoid of Trump supporters. I think the analysis is very interesting, mainly because of the breadth of data mined. I especially enjoyed the statistics and visuals on the issues which were most discussed.
And now, over to you: do the results and conclusions here ring true to your experiences following election news? Which platforms did you use most often to consume political news during the election, if any?
“How the Circle Line rogue train was caught with data” over at Data.gov.sg — Singapore’s open data portal — is a picture perfect example of “using data to solve real world problems.” It makes for a fun and interesting read, and I like that they provide the code used for their analysis. The outcome of the analysis solved a problem for thousands of daily commuters.
The world is better than you think, and it’s getting better (in many, many aspects). Despite local tragedies that reach enormous scales (such as in Syria right now), the data shows that the world is constantly getting better in many ways. In “A history of global living conditions in 5 charts” at Our World in Data, each chart allows data exploration with smart interactivity. Can you provide reliable data sources that might support or contradict these findings? Please share them in the comments.
The Conference and Workshop on Neural Information Processing Systems (NIPS) is a conference on machine learning that had well over 5000 participants this year. I did some digging to learn more about the conference’s most popular topics.
Corey Chivers’ post, “Generative Adversarial Networks are the hotness at NIPS 2016,” illustrates what Generative Adversarial Networks do. I also enjoyed Victoria Krakovna’s AI Safety Highlights from NIPS 2016. She’s a Google DeepMind research scientist who is interested in artificial intelligence safety. You’ll also want to check out this review of Andrew Ng’s lecture “The nuts and bolts of Applying Deep Learning.” Lastly, you don’t want to miss this extensive and informative review of the NIPS conference on Aylien’s blog. Aylien is an artificial intelligence, natural language processing, and machine learning startup based in Dublin, Ireland.
At The Guardian, I appreciated Hannah Devlin’s piece on bias in machine learning algorithms, “Discrimination by algorithm: scientists devise test to detect AI bias.” I’m very interested in whether algorithms are fair and unbiased, and how to find out if they are not.