• Machine Learning in Clinical Trials and R or Python or Both?

    Thanks to this event's sponsors IQVIA (https://www.iqvia.com/) and Linode (https://www.linode.com) for their generous support of DataPhilly! We couldn't make DataPhilly happen without their help. If you're interested in sponsoring future events please fill out our form at https://goo.gl/JLVfqh. We present two excellent talks on Data Science. * Nan Li from IQVIA will talk on using deep learning techniques to find the most appropriate clinical investigator. * Scott Jackson from Rstudio will address the debate of R vs Python by demonstrating the need for both. ---------------------------------------------------------------------------------------------------- Dr. Right: Optimizing Clinical Trials with the Most Appropriate Investigators The success of clinical trials heavily depends on choosing the top-enrolling investigators. We develop a recommender system for ranking investigators based on their expected enrollment performance on new clinical trials. I will be sharing how we integrate heterogeneous data, including the free-form text protocol data of clinical trials, investigators’ historical enrolling performance data, EHR data and so on, and our novel algorithms based on advances in deep learning and tensor mining. **** About Nan Li **** Nan Li is a Data Science lead at IQVIA, where he builds machine learning models to optimize clinical trials. He is passionate about solving human data science problems with Artificial Intelligence. He has both AI research and software development backgrounds. ------------------------------------------------------------------------------------------------------ Using R and Python Together Effectively "Should I learn R or Python?" is a question you'll hear frequently from beginner data scientists, who are eager to enter the field as rapidly as possible and focus on a unified toolset. Many people follow the principle of the "best tool for the job" and believe that a data scientist should be eventually proficient in both. However, sometimes the best tool for the job is a combination of both languages. Scott will be discussing how you can benefit from both languages' strengths using Reticulate and Feather, how to think scientifically about your performance concerns using experimentation tools, and how you can successfully productionize polyglot data science projects. **** About Scott Jackson **** Scott Jackson is a software engineer at RStudio and individual contributor to RStudio Connect. Scott has been writing software professionally for over 14 years and continues to do so, despite a brief career as a patent lawyer. Scott lives in Center City Philadelphia, roasts coffee for fun, and wants to know about fascinating open source stuff you have done or are thinking about doing.

  • Save the Date...


    We are working on an exciting speaker event for this date.