Past Meetup

Predictive Modeling in the Cloud w/ BigML and yhat

Hosted by SF Data Science

Public group

This Meetup is past

207 people went

Location image of event venue

Details

We're super excited to share with you two great tools that allow you to build predictive models in the cloud: BigML and yhat.

Here's the schedule for the evening:

6:30pm-7pm: Doors open, food & mingling

7:00pm: Dave Gerster, VP of Data Science at BigML

7:40pm: Greg Lamp, Co-founder of yhat

8pm-9pm: Networking

>>You must submit your first and last name and bring ID in order to attend.

Yelp works hard to keep their office secure. Their security people + someone from Zipfian Academy will be checking IDs.

If your name on Meetup.com does not have your first and last name, you must submit it through this form:

https://docs.google.com/forms/d/13EDHR2aWQY260xuwAi3f4dpwJECdXwA7gSP9ZVgGG8M/viewform

Thanks!

Speakers:

Dave Gerster, VP of Data Science at BigML

Let’s say have a clean data set (finally!) and are ready for the fun part: analysis and model-building. There’s a lot of things you can do at that point.

One awesome option is to start your analysis with a decision tree in the cloud. On BigML, you can make one in just 10 minutes. There's also a powerful way built-in to explore confidence in its predictions.

We’re stoked to have Dave present. He is the VP of Data Science for BigML, was formerly the Director of Data Science at Groupon, Principal Program Manager at Bing, and Director of Analytics at Yahoo!. (Whew.)

Dave will show us through BigML and discuss how it works under the hood.

Greg Lamp, Co-founder of yhat

We’re also very excited to share with yhat with you. Think of their company as the Heroku of data science - they make it super-easy to deploy predictive analytics.

Once you've written your model in R or Python, you might hand it off to a production engineer to implement. However, this work is time-consuming and error-prone. According to Greg and his co-founder Austin, you can deploy your model in minutes on their platform.

The two of them are in town from New York for just a few days, so this is an extra-special opportunity to see them.

Thanks, Yelp!

Thanks very much to our host and sponsor of pizza and beer, Yelp!. Their events specialist Julia is fantastic, their space is beautiful, and they’re hiring! See: http://www.yelp.com/careers.

Get in touch with the organizers if you'd like to host a future event.

More about BigML

BigML offers a highly scalable, cloud based machine learning service that is easy to use, seamless to integrate and instantly actionable. Now everyone can implement data-driven decision making in their applications. BigML works with small and big data.

PS: Watch this video of Dave from BigML beating the benchmark for random forest models on the Kaggle StumbleUpon competition: https://www.youtube.com/watch?v=eU7ayE-U1eg

More about yhat

Integrating predictive models into production software is an incredible chore. Models written in one environment must be ported to that of the production system (e.g. Java, .NET, Ruby). Converting models from one language to another is impractical, error-prone, and time-consuming. Yhat lets data scientists publish predictive models in a manner that's immediately useful to application developers.

More about Yelp

Yelp (NYSE: YELP) connects people with great local businesses. Yelp was founded in San Francisco in July 2004. Since then, Yelp communities have taken root in major metros across the US, Canada, UK, Ireland, France, Germany, Austria, The Netherlands, Spain, Italy, Switzerland, Belgium, Australia, Sweden, Denmark, Norway, Finland, Singapore, Poland, Turkey, New Zealand, the Czech Republic and Brazil. Yelp had a monthly average of 108 million unique visitors in Q2 2013*. By the end of Q2 2013, Yelpers had written more than 42 million rich, local reviews, making Yelp the leading local guide for real word-of-mouth on everything from boutiques and mechanics to restaurants and dentists. Yelp's mobile application was used on 10.4 million unique mobile devices on a monthly average basis during Q[masked]