Author: DataScience

07Nov

11/08/2018 – Guest Speaker on ML and Data Mining in Python

Interested in Machine Learning & Data Mining (in Python)? 
On November 8th, we will be having a postdoctoral researcher coming in to speak about his work and some projects he’s completed. Dr. Kuusisto completed his Computer Science PhD in Machine Learning in 2015 at UW-Madison, and works in the Regenerative Biology lab at the Morgridge Institute to build models from genetic expression data that can predict when compounds are toxic to developing neurological tissues.
24Oct

Biostatistics at Vanderbilt University

The Graduate Program in Biostatistics at Vanderbilt University is seeking quantitatively oriented undergraduates (e.g., majors in statistics, mathematics, computer science, or quantitative sciences) who have an interest in pursuing a graduate degree in Biostatistics or Data Science. Our program has an emphasis on biomedical applications, statistical theory, and computational methods as well as a strong emphasis on traditional data science topics such as machine learning and computational algorithms. I have attached a copy of our program brochure. Please consider posting the brochure for students to see and forwarding on an electronic copy of this email to your students.

If you would like hard copies of our brochure, please contact our program manager, Amanda Harding at amanda.harding@vanderbilt.edu with your mailing address, and she would be happy to send them to you.

Vanderbilt DS/Biostats Brochure

24Oct

Talk: Leveraging Data Science in Health Care

Joe Tenini, a Data Scientist at Epic’s Inpatient Predictive Analytics R&D, will give this month’s DS3 Seminar. He will discuss the types of questions that his team is interested in, the massive data that they have access to, and some of his work that puts it all together.

The talk is public and should be accessible to anyone interested in data science. Join us! Please invite others! 10/28 at 3:30 in 140 Bardeen. Full details below:

Speaker:
Joe Tenini PhD, Data Scientist at Epic, Inpatient Predictive Analytics R&D.
When/where: 10/28 at 3:30 in 140 Bardeen.

Abstract: What would you do if you knew every medication administered, procedure performed, lab resulted, and diagnosis made for 190 million patients? What sort of questions could you ask? What sort of problems could you take on? What if you could deliver your insights directly to the patients and providers who need them?

For data scientists at Epic, these are questions we ask ourselves daily. In this talk we’ll discuss opportunities to put data to work in healthcare, the tools and technologies involved, and some specific challenges and solutions that come up during day to day work. This will be an interdisciplinary talk. Students and practitioners from all fields and experience levels are encouraged to attend and bring questions.

Bio: Joe Tenini joined Epic after receiving his PhD in mathematics from the University of Georgia. His current work centers on the modeling of patient deterioration and the development of early warning systems in the acute care setting.

If you are interested in connecting with Joe or the team at Epic, let us know!

Follow the event here: https://www.facebook.com/events/1608942902735871/

01Oct

Opportunities: Data Science and Basketball

An opportunity with the Milwaukee Bucks for a senior thesis! While the talk has past, feel free to follow up with Mike, Seth, or one of us to get more information:
Are you interested in writing a senior honors thesis using NBA data?
Mike and Seth (the analytics team at the Milwaukee Bucks) are open to sharing data and advice to students interested in NBA data projects.  If you are interested in pursuing this, please come to the talk today (info below).
As part of his talk, Mike will describe the types of data that are available.  Then, you will then need to submit a proposal for your project.
What should the proposal contain?
A proposal will likely contain these four elements:
(1) A focused question and a hypothesis.
(2) A description of the data that you will use.
(3) A rough description of how you would like to process the data.
(4) Preliminary thoughts on the types of analysis that will be performed and an idenficiation of key hurdles.
What makes a proposal great?
The proposal should clearly communicate the aims and methods.  The proposal should be focused and interesting.  If it is not obvious, it should explain why the proposed question is answerable with the available data.  The very best proposals use the publicly available data (there is a lot of it) to perform a preliminary analysis or a “feasibility study”.  Finally, the final product of the research should be useful for the team.
How will the proposal be judged?
Does the proposal clearly communicate the aims and methods?
Is it focused and interesting?
If it is not obvious, does it explain why the proposed question is answerable with the available data?
Is the final “product” of the research useful for the team?
The very best proposals use the publicly available data (there is a lot of it) to perform a preliminary analysis or a “feasibility study”.
How long should it be?
No more than 2 pages.  Shorter is better.
How do I submit the proposal?
Email a pdf to karlrohe@stat.wisc.edu and derekb@stat.wisc.edu by Nov 11.
About the Talk:
Abstract: In this presentation, we will discuss data science through the field of professional basketball. However, many of the topics covered will have wider applications. We will discuss our approach to basketball analysis using specific examples of data design, automation, and research. We will also discuss the importance of succinctly communicating the analysis and visualizing conclusions. Following the presentation, we will allow time for Q+A.