One of the highlights from Joseph Sirosh's keynote was a demonstration of the World Wide Telescope, an online tool that allows budding astronomers (and professionals!) to explore the visible universe using imagery from the Digitized Sky Survey. The Survey contains images from thousands of galaxies that have not yet been classified by astronomers into one of the standard forms: elliptical, spiral, or lenticular. In the demo, the World Wide Telescope automatically classifies the galaxies in mere seconds:
The automated classification was based on a random forests model, using the manually-classified galaxies in the database as the training set. The model was trained and run using R (specifically, SQL Server R Services) running within SQL Server 2016. You can see the R code embedded within the T-SQL used in SQL Server in the screenshot below.
Notably, R wasn't the only open-source techology featured in this demo. In fact, the demo was running SQL Server on Linux, which is in preview now and will be available in 2017.
If you'd like to explore R and SQL Server in more detail, the launch also saw the release of a number of in-depth videos featuring many of the developers involved in the project. (While I had a small cameo in the Scott Guthrie's keynote presentation on some of the team involved in the project, the names below deserve much more credit for the R integration.) Click on the links below for the videos:
- Scaling your advanced analytics in SQL Server 2016 with R Server: Bill Jacobs.
- R Services in SQL Server 2016: Dotan Elharrar and UC Jayachandran.
- R Services FAQ SQL Server 2016: Dotan Elharrar, Vijay Jayaseelan, Jasraj Dange, and UC Jayachandran.
- Building intelligent applications using SQL Server R services: an End to End Walkthrough, Gopi Kumar and Hang Zhang.
Official Microsoft Blog: SQL Server 2016: The database for mission-critical intelligence