NYU Data Science newsletter – December 1, 2015

GROUP CURATION: N/A

… Know your stats.

Math like linear algebra and calculus are more or less expected of anyone we’d hire as a data scientist, and we look for people who live and breathe probability and statistics. Promising candidates will have the equivalent of at least 3 or 4 courses in probability, statistics, or machine learning — anything beyond that is icing on the cake. You should be able to ace the homework and exams in your probability and stats courses — many of our data scientists have actually taught these courses before coming to Google.

Smaller tech cities do have a serious downside when it comes to the tech world: they often lack big-time venture capitalists and the pool of startup-savvy business and marketing talent that can help a small company grow.

[Luis] Von Ahn and other startup CEOs, though, are beginning to push back against this argument.

They make the case that Pittsburgh and other second-tier tech cities—including Raleigh, St. Louis, and Minneapolis—are places with strong university pipelines, affordable living costs, great quality of life, and collaborative tech ecosystems. “Despite the fact that a lot of people have told us to leave,” von Ahn says, “we’re happy here.”

… Yu Zheng is a lead researcher with Microsoft Research Asia, which is Microsoft’s main research facility in the Asia Pacific region that houses more than 250 researchers and developers, along with thousands of other employees.

Zheng and his colleagues focus on using big data and machine learning to tackle urban problems. One project they’ve worked on is called Urban Air, an interactive map that lets users see air quality levels across 72 cities in China.

In this day and age, you can easily share photos through Dropbox, notes in Evernote, or spreadsheets via Google Drive with anyone. But good luck helping two doctors at two different hospitals to see the same patient records online. Instead, when a patient goes to a medical center for the first time, they often have to repeat tests they’ve undergone before—such as a computerized tomography (CT) scan, which uses X-ray technology to produce cross-sectional images of the body.

“The holy grail of medical informatics right now is to have a cloud-based place where patients’ info can live,” says Dr. Alexander Baxter, an assistant professor of radiology at NYU who practices at Bellevue and NYU hospitals in Manhattan. “So that if you go to one hospital, and you get a CT scan and you go to another hospital, you don’t have to get the same CT scan again. This happens all the time at Bellevue.” That doubles the cost and the dose of radiation.

Big data is a buzzword commonly tossed around in industry these days. Most simply, big data is a type of statistics—a way of analyzing extremely large data sets with computer algorithms to reveal important patterns and correlations. Corporations in healthcare, retail, and other industry sectors have been trying to leverage big data to help their bottom line for years.

Now large-scale scientific projects are adopting the method, including the National Science Foundation as well as the US BRAIN Initiative. Francis Collins, director of the National Institutes of Health, says this important basic science undertaking—developing tools and technologies to help create a “dynamic understanding of brain function”—will be fueled, in part, by big data and analytics approaches.

Booz Allen Hamilton (NYSE:BAH), a leading provider of management consulting, technology and engineering services, today released its second edition of The Field Guide to Data Science.

This latest edition is an update to The Field Guide to Data Science, which was originally released in November of 2013. The resource is widely regarded throughout government and industry as the definitive guide to data science. It has served as the foundation for the definition and role of data science within major government and commercial organizations, and universities have integrated it as part of their data science course work.

I’ve recently been working on some simulation studies in R which involve computer intensive MCMC sampling. Ordinarily I would use my institution’s computing cluster to do these, making use of the large number of computer cores, but a temporary lack of availability of this led me to investigate using Amazon’s Web Services (AWS) system instead. In this post I’ll describe the steps I went through to get my simulations going in R.

A few years back my wife kindly hosted me at the McGill CS Graduate Student Seminar Series. It was a well attended, candid talk about how to succeed at incorporating natural language processing (NLP) techniques within graduate research projects outside NLP itself.

Given the nature of the talk, I did not find sharing the slides as I do for my other talks to be that useful. Instead I’m putting that content into this blog post.

NYC Media Lab has a unique opportunity to build on the community of faculty, startups and media executives and technologists it has created to build a ‘spinoff engine’ focused on commercializing media technologies from universities. The program is supported by funding from the New York City Economic Development Corporation, the Mayor’s Office of Media and Entertainment, and through the support of NYC Media Lab’s corporate membership.

The goal for the Combine is to match new technologies emerging in university labs, studios, classrooms and dorms with a program and resources to commercialize them. Key to the proposition is mentorship and market access from technologists and executives at NYC Media Lab member companies.

Deadline for Application: Tuesday, December 15

Call for Papers: Fourth Annual Workshop on Crowdsourcing and Online Behavioral Experiments (COBE 2016), a workshop at WWW 2016, Montreal, Canada
… The main purpose of this workshop is to bring together researchers conducting behavioral experiments online to share new results, methods and best practices.

Deadline for Submissions: Tuesday, December 22