Registration

Conference Tickets


Tutorials

Registration for tutorials is open to everyone who has purchased any type of ticket for the conference. After purchasing a ticket for the conference, please e-mail gro.tats-pu@ofni with the tutorials you'd like to attend. Seating capacity is not updated in real time; there is no guarantee that seats will be available by the time your request is processed so you may want to send an ordered list of the tutorials you'd like to attend. Requests are processed in the order in which they are received.

A Gentle Introduction to Statistical Learning Theory for Data Science by Ernest Fokoue 0 seats remaining
Friday 9:00 AM - 11:00 AM
Description:

Coming Soon
Human Factors in Graph Design by Esa M. Rantanen 28 seats remaining
Friday 9:00 AM - 11:00 AM
Description:

A graph, or a chart, is a graphical (i.e., visual) representation of data. The primary use of graphs is to visually and thus very concisely communicate information about relationships between different things (variables, system components, &c.). Don Norman's idea about three conceptual models, or the design model, the user's model, and the system model applies also to graph design. The design model is the conceptualization of what the researcher had in mind, what he or she deemed important to communicate it the reader. The user's model is what the reader makes of the graph and how he or she understands the system (variables and their relationships) presented. Ideally, the design and user's models should identical or close to identical. The system model is what the graph actually conveys.
Kaggle Predictive Analytics with Random Forests and Boosted Trees by Padraic Neville 12 seats remaining
Friday 1:00 PM - 3:00 PM
Description:

Kaggle is an online forum for predictive modeling competitions. Among the competitions that ended in 2015, 26 were open and free to everyone and gave prize money to the winners, with total outlays from $250 to $175,000, a median outlay of $18,000, and a median number of entrants of 667. Most participants become better modelers, and several competitions are designed specifically to help beginners learn how to do predictive modeling. The variety of competitions testifies to the ubiquitous need of predictive modelers. Kaggle is also a laboratory for the next generation of predictive methodology, and the new models look bizarre: big and complex beyond comprehension, typically devoid of insight into the data.

The competition winners sometimes explain how they analyze the data. This tutorial draws from these valuable testimonials to show how competitive predictive models are developed. Neural Networks and Deep Learning dominate competitions with image data, while tree-based models, especially xgboost, are preferred in the others. The first half of the tutorial will therefore introduce decision trees, random forests, gradient boosting, and xgboost, with discussions on overfitting, unbiased variable selection, and variable importance measures.

The second half explains the mechanics of participating, risks of overfitting the leader-board, and details of several winning strategies. Feature engineering is contrasted with autoencoding and embedding. Although most contestants rely on R or Python, this tutorial is introduces analytical ideas and does not require any knowledge of programming.
Analyzing Gravitational Wave Data From the LIGO Open Science Center by John Whelan 21 seats remaining
Friday 1:00 PM - 2:00 PM
Description:

On September 14, 2015, the Advanced LIGO Detectors in Livingston, LA and Hanford, WA observed a gravitational wave from a binary black hole merger, known as GW150914. Concurrent with the publication of the resulting scientific paper on February 11, 2016, the LIGO Scientific Collaboration released approximately one hour of data from the LIGO detectors around the time of the event. In this tutorial, I will show how to use these data and other resources to replicate analyses such as the matched filter used to establish the detection.
Practical Natural Language Processing by Emily Prud'hommeaux 19 seats remaining
Friday 2:00 PM - 3:00 PM
Description:

Coming Soon
Tidy Data Analysis in R with dplyr, ggplot2 and broom by David Robinson 17 seats remaining
Saturday 10:00 AM - 12:00 PM
Description:

Coming Soon

Important Dates

Abstract Submission Deadline
Friday, March 11, 2016
Notification of Acceptance of Submission
Friday, March 25, 2016
Data Competition Entry Deadline
Tuesday, March 15, 2016
Data Competition Submission Deadline
Wednesday, April 13, 2016
Online Ticket Sales Close
Thursday, April 21, 2016
Registration Deadline
Thursday, April 21, 2016
Organized Session Submission Deadline
Friday, March 4, 2016

Conference Topics

This year's conference will focus on (but is not limited to) Data Science, Statistical Practice, and Education. Submissions on the following topics are encouraged:
  • Novel contributions to statistical methods or computing
  • Applications of statistical methods to interesting data sets from biology/medicine, social sciences, business/finance, and other fields
  • Issues in statistics / data science education
  • Statistics education in secondary schools (and beyond)
  • Other aspects of statistical methodology and applications

When and Where


Canisius College
2001 Main St
Buffalo, NY 14208


Cocktail Hour & Poster Session
When: Friday, April 22, 2016 , 6 P.M. - 7 P.M.
Where: Grupp Fireside Lounge, 2nd Floor of Student Centre

Banquet
When: Friday, April 22, 2016, 7 P.M. - 9 P.M.
Where: Regis Room, 2nd Floor of Student Center

Conference
Friday, April 22, 2016
  • 9AM - 3PM: Tutorials
  • 3:15 PM - 4:15 PM: Fr. Haus Memorial Mathematics Lecture
  • 4:30 - 5:55 PM: Panel Discussion ("The Multiple Facets of Data Science")
Saturday, April 23, 2016
  • Conference: 8:00 A.M. - 5:30 P.M.

Contact

Abstract Submissions
gro.tats-pu@stcartsba
General Conference Questions
gro.tats-pu@ofni
Website Issues
gro.tats-pu@retsambew