About Myself

Investigator, Scientific Data Analysis	ScD, Biological Engineering, 2017	Bayesian Stats, ML, Network Science

I don't have much advice to offer, just a bunch of stories that hopefully entertain and educate.

Much of my journey into data science has been the culmination of:

• Being helped by lots of friendly people.
• Having sufficient time and space to grow intellectually.
• Getting lucky, many times.

In other words, survivor bias is present in my story.

What is a data scientist?

A data scientist is someone who uses data to make better decisions and ask better questions.

It just so happens that at this moment in time:

• Dealing with data correctly requires statistical thinking.
• Manipulating data at scale requires hacking with code.
• Gaining insights from data appropriately requires domain knowledge.

These are the skills that a data scientist probably needs at the moment.

NIBR>NX>SDA>MLQ

• NIBR: Novartis Institutes for Biomedical Research
• NX: NIBR Informatics
• SDA: Scientific Data Analysis
• MLQ: Machine Learning & Quantitative Analysis

Projects @ Work

• Bayesian simulation of NIBR's Project Portfolio.
• Analysis of high throughput RNA cleavage.
• Scaled, hierarchical Bayesian analysis of electrophysiology data.
• Systematic benchmarking of message passing neural networks.
• Analysis of live brain imaging videos.
• Initiative: differential computing toolkit.

My path to Data Science

A journey in 4 acts.

Act 1: Get a fresh start

Getting let go from my first lab gave me the space to:

• switch to a much more forgiving and safe environment.
• learn many new topics.

Started going to Boston Python User Group.

• Met new friends.
• Learned new Python tricks.

Act 2: Work on a data project

My thesis gave me the opportunity to work with data, at scale.

Getting my thesis formally peer reviewed and published in a journal was, for me, validation.

Actually, work on more than one data project.

Other projects during grad school

• Quantitative measurement of influenza binding by fluorescence imaging.
• Geospatial mapping of influenza movement with wild birds.
• Low-cost remote imaging of wild seals with Raspberry Pi cameras!
• Predict influenza virus host species from protein sequence.
• Predict drug resistance from protein sequence.
• Forecast viral evolution with generative models of evolutionary trajectories.

Act 3: Share Knowledge

...it is more blessed to receive than to give...

Acts 20:35

Some things I've learned and shared along the way

• Bayesian statistical analysis recipes, tutorial.
• Open source software contributions.
• Network analysis workshops.
• Talks on Bayesian deep learning and statistical inference.

Some side effects of sharing knowledge:

• Reinforcement learning. For humans.
• People will mistake you for an expert.
• Make many new friends along the way.

Act 4: Meet kind people, and be kind to others

Jon Runstadler, Nichola Hill, Islam Hussein

Doug Lauffenburger, Forest White

David Duvenaud, Matthew Johnson, Dougal Maclaurin

Giles Hall, Ned Batchelder, Carol Willing, Allen Downey, Travis Oliphant, Matthew Rocklin

Hugo Bowne-Anderson, Mridul Seth

Nick Baro, Andy Long, Lily Xu, Ivan Zheng, Greg Antell, Faye Zheng

What names are part of your professional story?

To whom can you be thankful for giving you a hand?

Parting Thoughts

Stay curious, don't stop learning.

Make friends.

Remember those who gave you a hand.

To whom much is given, much is expected.

Luke 12:48

Give more to society.

One of Novartis' corporate goals.