Dr. Keith O'Rourke talks about the logical reasoning behind statistical modeling. Topics include mathematical vs scientific reasoning, whether science has become too stats focused, and vice versa.

Watch it on... Youtube: https://youtu.be/FqE4ROHBKpY Podbean: https://dataandsciencepodcast.podbean.com/e/keith-o-rourke-the-logic-of-statistics/

Topic List:

0:00 - The logic of statistics 0:30 - What is scientific statistics? 5:15 - The logic of statistics and CS Pierce 9:15 - Role of representation in statistics: explicit vs implicit 14:13 - Diagrammatic Reasoning 18:45 - Why is modeling counterfactual? 19:33 - How can statisticians become better scientists? 28:40 - Science is hard 31:24 - Computational approaches to learning 42:00 - Learning through metaphor 46:28 - Diagrammatic representations vs math 48:40 - Is science too statistics-focussed?  59:35 - Is statistics sufficiently science-focussed?  1:08:40 - Scientific Debate

#statistics #datascience #science 

Did you know that it's possible to hide malware in neural networks? Actually, you can hide malware in many statistical models. This is the subject of two recently-published papers (aptly titled "EvilModel" & "EvilModel 2.0"). Dr. Jack Fitzsimons makes it easy to understand how this is done, using techniques that began long before computers.  

Watch or listen on...  Youtube: https://youtu.be/QBnk8ogL8Nk Podbean: https://dataandsciencepodcast.podbean.com/e/jack-fitzsimons-evil-models-hiding-malware-in-neural-networks/

Watch it on... YouTube: https://youtu.be/yNaCudDVTkY Podbean: https://dataandsciencepodcast.podbean.com/e/scott-cunningham-causal-inference-the-mixtape/

0:00 - COMING UP... 0:35 - What makes it into the mixed tape? 7:10 - Coding to learn 11:15 - More people are expected to work with data & code 12:50 - Design vs program vs estimators 20:40 - Causation with zero correlation 27:00 - Optimization make everything endogenous 28:45 - The hospital example 29:30 - Credible scientific discovery vs motivated discovery 39:55 - Different meanings of causality 43:30 - The impossible counterfactual  47:00 Counterfactual nihilism 49:20 Social experiments / Defund the police 53:35 - Skepticism about the science of social phenomena 1:05:20 - The Italian crime example 1:16:30 - Scientific debate

YouTube: https://youtu.be/K5nsSMJVIT0

Andrew Gelman and Aki Vehtari wrote a paper titled, "What are the most important statistical ideas of the past 50 years?". The first idea in the list is "counterfactual causal inference". Eric Daza (Evidation Health) walks us through the main ideas of the Gelman & Vehtari paper, drawing examples from several fields, including medical & healthcare statistics. 

Topics 0:00 - Coming up...Correlation vs Causation 1:20 - Most important statistical ideas over the last 50 years 6:10 - Counterfactual Causal Inference 9:40 - Assumptions Change between Applied Domains 21:10 - Propensity Score Methods 25:15 - Transportability of Scientific Results  26:30 - People don't want generalizable results 32:00 - Generic Computation Algorithms 37:00 - Reweighting 43:57 - Matching Methods 58:20 - Medical Data is Higher Dimensional that we think. 1:00:15 - Is a Trial Population Representative?  1:10:35 - Causal Models in the Future 1:18:45 - Apostates Welcome 1:21:45 - Scientific Debate

Ruda Zhang (Duke University) walks us through "Gaussian Process Subspace Regression for Model Reduction" by Zhang, Mak, and Dunson.

To keep the topic interesting for both the early career & advanced audience we recap key points at a high level so that no one gets lost.

This episode involves a presentation, so you may prefer to watch the YouTube version here: https://youtu.be/IPtqUUG4XcY

Ruda's website: https://ruda.city/ The paper: https://arxiv.org/abs/2107.04668

Watch it on... Youtube: https://youtu.be/GoDwen-RGZg Podbean: https://dataandsciencepodcast.podbean.com/e/ruda-zhang-math-science-duality/

Statistics is thought to reside at the interface of science and mathematics. Ruda Zhang (Duke University) discusses the friction at this interface and the role that both mathematical formalism & observational/data-driven intuition play in scientific discovery. A great topic for anyone interested in statistics' role in scientific discovery.

#datascience #ai #science #mathematics

Topic List 00:00 COMING UP... 2:44 Ruda Zhang's compendium of cool ideas + a Gaussian process PSA 7:08 Is intuition undervalued in scientific research? 10:16 Mathematics vs observational science. Rigor vs intuition. 14:07 Intuition & discovery precedes mathematical rigor 21:58 Mathematics vs empirical science & the complexity of induction 30:24 Abstract thinking & the cost/benefit of discovery 37:25 The efficient frontier / Pareto Front of knowledge 42:55 Pragmatism and competence 50:24 Math /science dualism 1:15:52 AI making scientific discoveries 1:19:15 Statistical & scientific debate

It’s a common dictum that statisticians need to incorporate domain knowledge into their modeling and the interpretation of their results. But how deeply can scientific principles be embedded into statistical models? Prof. Simon Mak (Duke University) is pushing this idea to the limit by integrating fundamental physics, physiology, and biology into both the models and model inference. This includes Simon’s joint work with Profs. David Dunson and Ruda Zhang (also of Duke University).

Scientific reasoning AND stats. What more could we ask for?

Enjoy!

Watch it on....

YouTube: https://youtu.be/bUbZO7R4z40

Podbean: https://dataandsciencepodcast.podbean.com/e/simon-mak-integrating-science-into-stats-models/

00:00 - COMING UP….Scientists & Statisticians 02:09 - Introduction - Integrating scientific knowledge into AI/ML 06:08 - How much domain knowledge is sufficient? 09:15 - Choosing which prior knowledge to integrate into a model 14:49 - Black box & gray box optimization 19:50 - Non-physics examples of integrating scientific theory into ML models 22:45 - Scientific principles & modeling at different scales 27:20 - Correlation is one just way of modeling linkage 36:37 - Conditional independence & different-fidelity experiments 39:40 - Innovation vs incorporation of known information in the model 42:52 - Aortic stenosis example 52:49 - Which mathematics can be used to represent scientific knowledge 57:09 - How to acquire scientific domain knowledge 1:02:45 - Complementary approaches to integrating science 1:06:48 - Gaussian process & integrating priors over functions 1:12:48 - A topic for statisticians and scientists to debate:science-based vs data-based learning.

Simon Mak's Webpage: https://sites.google.com/view/simonmak/home

#ai #datascience #startups

Martin Goodson (Evolution AI) describes the key aspects of the UK's AI Roadmap & responses to the document by members of the Royal Statistical Society. In particular, Martin describes the disconnect between the priorities of AI startups and industry practitioners on one side, and government and academia on the other. Martin also outlines which skills early career data scientists should focus on while in school versus after entering the workforce.

Also available on....

YouTube: https://youtu.be/T9qRl6Hclhg

Topic List

0:00 COMING UP: Scientific culture & AI

1:25 The UK AI Roadmap

8:44 Who is a data science “practitioner”? 

12:53 Data science in AI startups

20:36 Is there a disconnect between practitioners & academia?

25:09 Key skills for new data science graduates

32:03 Coding & production level data science

39:30 Learning the right data analysis skills at the course-level. 

45:32 AI leadership

58:40 AI from academia & OpenSource initiatives

1:05:37 Large institutions' impact on the AI field

1:08:24 Back to the UK AI roadmap  

1:12:16 Building an AI community 

1:13:15 AI in our lifetime: Moonshots & realistic goals

1:14:31 Scientific debate

#datascience #privacy #ai

0:00 - Coming up... 3:24 - Introduction 6:20 - Data privacy and privacy enhancing technologies   13:00 - History of privacy enhancing technologies 19:54 - Differential privacy: Hiding the influence of a single data point 22:52 - Trading data utility for data privacy 38:32 - Tracking algorithms and how they decide user preferences 42:04 - Preserving privacy: Anonymizing data & VPNs 50:17 - Exploration vs Exploitation: Combining best of multiple domains to tackle problems 54:13 - Federated learning, input and output privacy of data 58:45 - Balancing data privacy vs data-driven personalization 1:05:50 - What should data scientists/statisticians debate?

If you enjoyed the science-vs-pseudoscience topics, you’ll enjoy this one.

0:00 - Coming up in the episode

2:35 - What is the Piranha Problem?

19:54 - Confusing effect sizes

23:11 - The "words & walking speed" study

26:22 - Declaration of independent variables

30:58 - Piranha theorems for correlations

37:07 - Piranha theorems for linear regression

40:37 - Piranha Theorems for mutual information 

44:13 - Bounds on the independence of the covariates

46:12 - Applying the piranha theorem to real data

50:12 - Applying the piranha theorem across studies

54:05 - A Bayesian detour

1:00:12 - The butterfly effect & chaos

1:04:26 - Applying the piranha theorem to cancer research

0:00 - Intro 1:38 - Chris Holmes, Professor of Biostatistics at Oxford University 3:28 - UK Biobank & designing a valuable dataset 8:42 - Healthcare charities in the UK 11:16 - Digital Health: prioritizing research questions 19:55 - Bayes, nonparametrics, and Bayesian nonparametrics 23:30 - Model prediction is at the heart of Bayesian inference 28:00 - Prioritization in model building for biology 33:09 - Model constraints to generate valid inference 37:34 - Hypothesis driven science in statistical learning versus deep learning 43:30 - Developing models in genomics & clinical informatics 48:37 - Building stable, generalizable and robust models 52:41 - Important questions to think about  54:05 - Causal reasoning and clinical risk prediction 57:50 - What topic should the statistical community debate?

In the first keynote of the Philosophy of Data Science Series we have a 2-part interview with Deborah Mayo (Virginia Tech). In the first part of our keynote with Deborah Mayo we cover... - The role of scientific revolution and its implications for statistics and data scientist. - The necessity of statistical reforms and why philosophy will play a role. - The value of severe testing of scientific claims.

Watch it on...  YouTube: https://youtu.be/S4VAEShM3BU Podbean: 

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

Topics:

0:00 - Preface to First Keynote Interview 2:00 - Welcome Deborah Mayo! 5:05 - What is the Philosophy of Statistics? 8:15 - What does philosophy add to data science? 16:10 - Scientific revolution in statistics 20:10 - Statistical reforms 24:25 - Replication & hypothesis pre-specification 31:00 - Failure is severe testing 37:25 - Error statistics 48:00 - Scientific progress and closing remarks

Charlotte Deane (Oxford University) talks about statistical approaches to bioinformatics, the evolution of Google Deepmind's AlphaFold 2 & its place in protein informatics deep learning landscape. She also describes humanizing antibodies, and the increasing role of software engineers in statistical research groups. The topic of llamas, camels, and alpacas (and their unique place in proteomics research) makes a surprise visit.

[Note: This episode was originally published in January 2022, but the file contained a buffering error, which prevented the full interview from being played. This version, published Feb 1, 2022 contains the full interview.]

Topics 0:00 Intro / An important topic to debate 3:50 What is a protein? Why are proteins foundational? 13:32 Immunotherapies, humanizing antibodies, & creating an scientific databases 16:04 Translating in silico research into immunotherapies 21:03 Nanobodies, camels, alpacas, & llamas.  25:05:00 Databases and data knowledge bases 33:21:00 Targeted therapies 39:45:00 Statistical modeling in proteomics 45:40:00 DeepMind AlphaFold's evolution 55:28:00 Software engineers in academic research groups 1:03:21 The adventure of science 1:07:42 Oxford Blues hockey & scientific debate

Keywords: philosophy, epistemic data, first person data, stimulus error, imageless thought, consciousness

The following interview was a keynote fireside chat with Janet Wittes (Statistics Collaborative, Inc.) titled "Statisticians as Entrepreneurs". It was recorded for the BBSW 2021 Conference (Nov 3 - 5 in Foster City, CA).

References:

BBSW 2021 Conference: https://www.bbsw.org/bbsw2021

Topics:

0:00 Janet's background prior to founding Statistics Collaborative, Inc. 3:00 Janet's initial research interest as a consultant 4:10 Why did Janet start her own business as opposed to joining a company or university.  5:45 Who were Janet's first clients? 8:00 What did Janet want to instill in her company? 15:50 Earning enough money to hire people 18:55 Initial ratio of clients to employees 22:42 Janet's company's statistical tech stack 25:00 Different challenges at different stages of the company 27:28 Growing a company but not taking on every possible client or project 28:13 Statisticians as entrepreneurs 37:00 Choosing the right people

Watch it on….    YouTube       Podbean

Jingyi Jessica Li (UCLA) describes common statistical pitfalls in genomic data analysis & the statistical reasoning required to correct these mistakes.

Common themes throughout include:

• Hypothesis-driven science & critical scientific reasoning over data
• p-values and non-sensical null hypotheses/distributions
• the value of appearing statistically rigorous
• researchers cutting intellectual corners & digging themselves into local minima

Episode Topics

0:00 A major advancement in genomic data leads to new statistical techniques

2:15 Hypothesis-driven science & hypothesis-free data analysis

2:55 A ChIP Seq Example

8:00 Misformulation of sampling variability

16:55 A false analogy: the permutation test

19:03 Losing my p-value religion: the value of statistical packaging

24:30 The Clipper Framework for false discovery rate control

31:50 Non-parametric developments

37:55 Inferred covariates

46:00 PseudotimeDE: inferences of differential gene expression along cell pseudotime

47:10 Selective inference

49:25 What biological/physiological data will be incorporated in the future?

52:30 Statistics, computer science, data science, ML, biology

57:05 Machine learning and prediction

1:01:30 Sophisticated models vs sophisticated research

1:07:45 Peer review in science

1:13:05 Hypothesis-driven science vs cutting intellectual corners

1:18:12 What topic should the statistics community debate?

Mine Çetinkaya-Rundel (Duke University) describes the current and future states of statistics and data science education. Then she discusses the process of building open access learning material.

0:00 - Introduction 1:40 - Prioritizing topics in curricula 9:07 - Teaching with intent to test 11:22 - Statistics without computing 17:52 - What should be taught? How do we teach it? 19:07 - Computational thinking is valuable (to 31:45) 23:47 - Self reinforcing academics / positive feedback (to 31:45) 31:08 - Data science vs statistics (the computing angle) 37:55 - Statistical collaboration / technical collaboration 39:45 - Common language / imputation under ignorance 41:12 - Are some topics better for hands on or computational learning? 45:32 - Learning computation through visualization 52:40 - Video cut option before she gives an example 52:42 - Let them eat cake first. 56:08 - What is open source education? Open source vs open access. 59:36 - Advancing open source text books 1:03:55 - Economics of open source 1:07:55 - The open education ecosystem 1:12:17 - Modularizing & parallelizing learning topics 1:16:52 - Favorite dataset on OpenIntro.Org? 1:18:14 - What topic should the statistics community debate?

Jingyi Jessica Li  (UCLA) discusses her paper "Statistical Hypothesis Testing versus Machine Learning Binary Classification". Jingyi noticed several high-impact cancer research papers using multiple hypothesis testing for binary classification problems. Concerned that these papers had no guarantee on their claimed false discovery rates, Jingyi wrote a perspective article about clarifying hypothesis testing and binary classification to scientists.

#datascience #science #statistics

0:00 – Intro 1:50 – Motivation for Jingyi's article 3:22 – Jingyi's four concepts under hypothesis testing and binary classification 8:15 – Restatement of concepts 12:25 – Emulating methods from other publications 13:10 – Classification vs hypothesis test: features vs instances 21:55 - Single vs multiple instances 23:55 - Correlations vs causation 24:30 - Jingyi’s Second and Third Guidelines 30:35 - Jingyi’s Fourth Guideline 36:15 - Jingyi’s Fifth Guideline 39:15 – Logistic regression: An inference method & a classification method 42:15 – Utility for students 44:25 – Navigating the multiple comparisons problem (again!) 51:25 – Right side, show bio-arxiv paper

First-person methods (and its associated data) have been scientifically and philosophically contentious. Are they pseudoscientific? Or simply pushing the bounds of scientific methodology? Obviously, I have no idea… so Prof. Gualtiero Piccinini (University of Missouri – St. Louis) provides a helpful introduction to the topic covering the key points of its history and the philosophical/scientific debate.

0:00 Why cover first-person methods & data? 2:26 First-person methods vs first-person data? 7:10 Are first-person data legitimate at all? 11:50 Phenomenology 13:26 First-person data is extracted from human behavior 18:25 Skepticism & arguments against first-person data 25:40 Psychophysics, introspectionists, behavioralists, cognitivists, and the origins of first-person data 35:20 Using new instruments & methods in science 46:00 Is this where the philosophers roam?

#datascience #statistics #science

A fundamental question in the philosophy of science is "what does it mean to make scientific progress?" We will have a series of episodes centered around this question for statistics and data science. In our first episode in the series, David Dunson (Duke University) discusses important advances in Bayesian analysis, big data,  uncertainty, and scientific discovery. 

Topic Timestamps 0:00 Intro to David Dunson 1:54 What does it mean to advance data science and statistics?  6:14 Industry & Optimization, Science & Uncertainty 8:14 Prediction & Discovery / Bayesian Modeling  14:13 What is “complex” data? 22:49 Big Data, Bayes, and Nonparametrics 33:50 Ad hoc approaches vs principled methods 37:08 Should Machine Learning Publications Refocus on Scientific Discovery? 39:50 Mathematically principled data science & statistics 51:40 Do Bayesians just use priors as regularizers? 55:16 Bayesian Priors and Tuning Inference Methods 1:00:00 Prioritize the Most Important Work in Data Science  1:07:07 Good Practices of Star Grad Students 1:13:17 The Science in Statistical *Science*

#datascience #science #statistics

Martin Kuldorff | Spatiotemporal Models of Outbreaks Martin Kuldorff (Harvard Medical School) talks about the integration of biological & demographic information (and general reality) in the spatiotemporal models used to detect disease outbreaks. He also discusses how these methods can be applied to non-infectious diseases like cancer.

0:00 - Spatio-temporal modeling of outbreaks 6:02 - Important features of spatio-temporal outbreak models 12:20 - Which diseases wouldn't you track for modeling? 19:02 - Multiple comparison adjustments of alarms 25:15 - Domain knowledge of outbreak features 29:30 Competing hazards & risks  34:30 Comparing hemispheres 37:00 - Bridging the gap for infectious diseases to cancer 45:10 - Retrospective data correction / changing monitoring  57:00 - Competing risks & statistics 1:01:30 - Deducing risks & affects through knowledge of immunological mechanisms 1:09:00 - Future scientific convos

#datascience #science

Jason Costello | Data Science vs Software, Academia vs Industry Jason Costello (Hypervector) describes his (non-trivial) transition from academic research into big tech and then the healthcare industry. He outlines a strategy to find the cool research problems that you get in academia while still delivering value to your company. We then talk about the interface of data science / machine learning and software.

0:00              Deploying Data Science into the Real World 8:24              Transitioning from Academic to Industrial Data Science 16:56            First step to delivering value to industry 21:38            Toy example of high value data science 25:28            Deep technical challenges are real and useful too! 29:59            Formalized logic in machine learning solutions 32:54            Data Science & Machine Learning Projects can fail. 38:50            Getting to the cool data science projects 47:21            Putting Machine Learning Models into Software 56:21            Software and Deduction, Machine Learning and Induction 1:06:06         Is Software A Deductive Complex System?

Eric Daza | N-of-1 Science & Causal Inference | Philosophy of Data Science

Much of our scientific inference revolves around the identification and replication of patterns in data. So what can be done when N=1? Eric Daza gives us a statistician's perspective on the ideas behind N-of-1 studies, its best examples, and strongest critiques.

0:00 - The purpose of N-of-1 & generalizability

3:30 - Successes and challenges in N-of-1

9:30 - A lightbulb moment

18:00 – Anomalies, Compliance, & Recurring Patterns

23:00 – Best Critiques of N-of-1, Safety, Efficacy

41:20 - Causal Inference

54:30 – Increasing the number of data scientists

1:03:30 – Biostatistics’ changing place in data science / statistical thinking

Edward McFowland III | Anomalous Pattern Detection & Model Building

Edward McFowland III (Harvard Business School) describes the differences between "anomalies" and "anomalous patterns". Edward describes how this informs modeling strategies, in particular, when to use an off-the-shelf model versus building a bespoke model from scratch. He then covers how to draw inspiration from different scientific and technical fields.

0:00 Edward: Live in Conference

2:00 Outliers vs Anomalies vs Anomalous Patterns

9:30 Strategy to Identify Anomalous Data Patterns

19:15 Adding Complexity to Models

25:00 Building Blocks vs Comprehensive Models

39:05 New Pieces of Evidence

40:40 Deciding Data Science Strategies

52:30 Connecting the Technical Dots

58:40 Interdisciplinary Interests

The Statistical Consulting Section of the ASA invited me to give a presentation on the data science job search followed by a Q&A.

They were kind enough to let me post it here (with minor edits).

My drawing of "cumulative cost" is wrong. It should intercept the "current cost" line at time = 0.

0:00 – Humility, Goals, & Human Data Points 5:00 – Play the Numbers Game 12:40 – Job vs Career 18:18 – Nonsensical Data Science Job Descriptions 25:40 – Technical Review & Presentation 30:00 – The Advantages of Early Career 37:25 – Save Job Descriptions / Industry vs Academia 46:10 – Career vs Job Clarification  53:10 – Bachelor’s vs Master’s vs Doctorate? 56:10 – Delivering Value Over Time 1:08:10 – Product vs Service  1:11:10 – Comments From an Academic Perspective 1:116:43 – Get Your Foot in the Door / Doing What You Love 1:25:50 – Future Q&A’s

Mike Evans (University of Toronto) describes his approach to statistical reasoning. Mike outlines how to recognize and address problems that are statistical in nature and why these approaches should be grounded in our ability to measure statistical evidence. 

Watch it on YouTube at: https://youtu.be/Q7JpGZxHxXU

0:00 Statistical Reasoning 2:30 The Basic Problem: Reasoning on Statistical Problems 13:00 Rules of Statistical Inference 19:30 Bias (The Controversial Bit?!?!) 24:10 Steps of Statistical Reasoning 25:50 Connection to Philosophy of Science 27:35 Measuring Evidence (Frequentist vs Bayesian vs Loss Function) 29:49 Problems with the p-values 32:00 Choosing & Checking Priors 49:25 Idealism, Good Plans, Bad Plans 54:45 Describing Your Reasoning 59:20 Critiques of the Principle of Evidence 1:04:00 Data-Driven Science vs Hypothesis Driven Science

Watch it on…       YouTube        Podbean

In our fourth episode of the “science vs pseudoscience” mini-series, Deborah Mayo (Virginia Tech) specifies several necessary criteria to be scientifically rigorous. She gives several examples of how statistical thinking is essential to scientific thinking and why she believes that the “I’ll know it when I see it” approach to delineating science from pseudoscience is not a good approach. 

Looking to catch up with the earlier “Science vs Pseudoscience” episode?

You can watch them here:      Intro Episode 1 Episode 2 Episode 3    

Keywords: biomechanics, models, metrics, gait, engineering, statistics, cognitive impairment, physical impairment

0:00 - Intro

03:01 - Creating models for performance optimization

07:23 - Why gait is an effective diagnostic tool

11:38 - Maximizing gait in creating models for post-ACLR

17:35 - Manifestation of different injuries & models

22:01 - Modeling motor control

26:28 - Applying other models in biomechanics

30:50 - Using asymmetric walking for recovery

39:30 - Understanding cognitive impairment recovery

44:19 - Moving forward with gait as diagnostic tool

45:40 - Taking inspiration from other fields / Statistics in Engineering

47:45 - Engineering and statistics hand in hand

52:50 - Limitations of modeling in biomechanics

54:20 - Starting a career in biomechanics

58:20 - Including cognitive impairment

1:00:20 - Tailoring models to specific cases

1:05:33 - Applying the models to injuries other than ACL

Michael McRoberts (Championship Analytics Inc.) uses Monte Carlo simulations to provide strategy analytics to college and NFL football teams. Topics include communicating data-driven recommendations, the need to create counterfactual data, and asymmetric decision rewards.

0:00 The challenge of sports analytics

5:00 Analytics recommendations

16:00 Communicating data-driven recommendations

24:35 Vegas Odds & Ancillary Data

30:00 Football is way behind / Data science projects with a "runway"

41:25 Creating experiments and counterfactuals

49:30 Implementing data science insights

56:15 Asymmetric decision rewards

58:50 How to start in sports analytics

1:10:00 Data science vs analytics vs statistics

#datascience #statistics #science #pseudoscience

Our science vs pseudoscience discussion continues with Andrew Gelman (Columbia) and Megan Higgs (Critical Inference LLC). Andrew and Megan describe two critical roles that statistics plays in science.... but also how statistics can add the air of scientific rigor to bad research or help statisticians fool themselves. From there the conversation goes on in a way that only a conversation with Andrew and Megan can! A very fun episode.

0:00 - Two roles of statistics in science 4:50 - Many models were intended for designed experiments 10:30 - The biggest scientific error of the past 20 years 15:00 - Feedback loop of over-confidence / Armstrong Principle 21:00 - Science is personal 25:00 - The value of different approaches / Don Rubin Story 34:40 - Statistics is the science of defaults / engineering new methods 45:00 - The value of writing what you did 52:27 - Math vs science backgrounds + a thought experiment 1:01:20 - Fooling ourselves

Please forward to any students or colleagues who would find this of interest!

We each like to think of ourself as scientific. I'm yet to meet someone who would embrace being called "pseudoscientific". But what makes the difference? In this episode, Neil Manson talks about the fallout from Thomas Kuhn's 1962 book "The Structure of Scientific Revolutions" and how this created a playbook for many modern critiques/attacks on scientific activity.

We have a new series that centers on the discussion of science vs. pseudoscience. Guests of different backgrounds share their insights on what really constitutes science and the highly-contested pseudoscience.  The implications for data scientists and statisticians is very interesting, since many of the examples around this debate involved the conflicts between hypothesis-driven science vs data-driven science.

0:00 - Intro 0:43 - Science vs Pseudo/Bad/No Science 05:52 - Demarcation problem of science 12:07 - Incentives in science 13:00 - Glen forgets the word for "book" 13:40 - Luminiferous aether & lunch tables 18:19 - Keeping “good science” out of the science category 22:49 - Aiming to define science in relation to Kuhn’s theory 29:06 - Kuhn’s theory in action in various scenarios 32:53 - Logical fallacies in the world of science 46:17 - Intelligent design theory as science 51:14 - Distinction of different sciences

00:03 - Introductions 5:33 - What is pseudoscience? 08:53 - Legitimacy of other sciences 12:11 - What qualifies as science? 19:00 - Inductivism and empirical falsifiability 26:22 - Positivism and the importance of assumptions 31:36 - Assumptions and observations for data scientists 42:34 - The pursuit of science 49:17 - Scientism and revolutionary scientists 54:43 - Pinning down what science is

Some fun reading on pseudoscience: https://philpapers.org/archive/MONP-1...

Thinking about a career change to Environmental Data Science? We invite you to listen to some career growth strategies and opportunities in environmental data science” podcast.

Throughout the episode we discuss how to transition from other careers to an environmental data scientist. How to get quantitative skills in order to switch to environmental science. Ways someone can learn environmental science and get an entry job as an environmental scientist. Plus, “in career growth should one focus on a specific domain or to go broad?”.

00:00:00 Start

00:03:14 Introduction

00:09:03 Transition from other fields to Environmental Scientist.

00:22:27 How to get quantitative skills in order to switch to environmental science.

00:36:20 In career growth should one focus on a specific domain or be broad.

00:44:00 Ways someone can learn environmental science.

00:48:49 Ways people can get an entry job as an environmental scientist

01:07:54 Final comments

Data Science Career Q&A for Undergrads with Mallory LaRusso

We continue to answer data science career questions. We've heard back from a lot of different groups about the world of data science. In this episode, we're talking about undergraduate DS job prospects. Mallory LaRusso is a senior at NCSU finishing her BS in Statistics, and Minor in Genetics. Watch/Listen as Glen and Richard answer questions from our guest Malory as she tries to understand ways of how to properly transition from being an undergrad student to becoming a data scientist. From questions about data scientists’ typical workday to their most challenging projects to date, we’ve got it all covered in this episode!

Keywords: data analytics, data science, programming, coding, workday, work culture, educational background

0:00 - Introduction

02:20 - Series overview

06:44 - Educational and career path

09:16 - Typical work day

17:00 - The importance of writing in data science

20:39 - Work culture

23:30 - Type of data you work with

26:00 - Mathematical vs Statistical Models

31:45 - The harder DS jobs are what’s left

32:35 - Favorite project as data scientist

36:00 - Work on a real problem 

39:55 - Data scientists’ degrees

44:15 - Difference of data analytics and data science

54:13 - Favorite programming language

1:01:29 - How data science jobs will change

1:07:42 - Largest data set to have worked with

1:17:02 - Advice for students to prepare for data science roles

1:36:30 - What advantage does an undergraduate have?

1:40:22 - Wrap-up

Philosophy of Data Science Series

Session 3: Data Science Highlight Reel

Episode 4: Alex Bolton on Step-change and Anomaly Detection

Who makes it into the highlight reel of data science? Alex Bolton for doing the hard work of analyzing data to figure out exactly when things don't look "normal". We discuss the critical reasoning behind step-change detection and anomaly/novelty detection. Alex provides several real-world examples of the data and challenges.

Watch it on...

YouTube: https://www.youtube.com/watch?v=097FO1JDkhU

Podbean: 

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

Session 3: Data Science Highlight Reel

Episode 2: Irina Gaynanova on Replicating Clinical Metrics & Innovating New Methods

Who makes it into the highlight reel of data science? Irina Gaynanova for her work on replicating clinical metrics for deployment. She then goes into how her grasp of the scientific domain helps her innovate new methods and metrics. Regardless of whether you work in the clinical domain, this is an example of rigorous scientific thinking in data science.

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

Watch it on...  YouTube: https://www.youtube.com/watch?v=DOf2lVHzZS4 Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-s3-e1-neuralnets-gans-causality-and-medicine/

Who makes it into the highlight reel of data science? Adler Perotte, because he's a clear thinker on why his data needs a specific type of analysis. In this case, it's the need to draw causal inferences from observational data. Go, GANS! Go!

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

Watch it on... YouTube: https://youtu.be/ftikMj7MoYM Podbean: https://podofasclepius.podbean.com/e/career-qa-10-questions-from-a-beginner-data-scientist/

This week's episode is likely of interest to early career data scientists or those interested in joining the field. Richard Franzese (Certara) & Glen Wright Colopy (Pod of Asclepius) team up to answer 10 questions from Ujjwal Oli, an MSc student at George Washington University MSc Program.

The questions range from technical requirements, to desirable soft skills and domain knowledge, to "how can I get an internship if they require prior experience?"

Please forward to any early-career statisticians or data scientists who would be interested.

Thank you for your support of the series! You can join the mail list here: https://www.podofasclepius.com/mail-list

#datascience #career #job #jobadvice

Philosophy of Data Science Series  Keynote with Deborah Mayo Episode 2: The Philosophy of Science & Statistics

In the first keynote of the Philosophy of Data Science Series we have a 2-part interview with Deborah Mayo (Virginia Tech). In the second part of our keynote, Deborah Mayo covers the interplay between scientific and statistical philosophy. Deborah highlights some common scientific fallacies, along with suggestions of where statistical thinking can be made more rigorous.

Watch it on...  YouTube: https://youtu.be/9GGAXZ6htrA Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-keynote-1-presentation-philosophy-of-science-statistics/

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

Session 1: Scientific Reasoning for Practical Data Science

Episode 4: Values and Subjectivity in Data Science

The Value-Free Ideal is a central tenant of objective science. But how do values, value judgements, and subjectivity leak into the practice of data science and statistics. To what extent is it desirable for science to be informed by values? Kevin Zollman (Carnegie Mellon University) covers the range of key ideas, from Heather E. Douglas to W.E.B. du Bois.

Watch it on...

YouTube: https://youtu.be/9USkWtX-ydc

Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-s01-e04-values-and-subjectivity-in-data-science/

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

0:00 Intro

0:03 Welcome Kevin Zollman (Carnegie Mellon University)!

1:44 Is Science Value-Free?

6:08 How might values affect science?

9:00 Choice of Research Problem

10:45 Loss Functions

18:34 Choice of Variables

24:10 Choice of Statistical Model

29:30 Minimizing the Values in Science (W.E.B. du Bois)

35:20 Philosopher in Science

41:20 Statements on Generalizability

47:45 Clarifying Subjective Choices

52:45 Conflicts between Scientific Disciplines

61:18 Scientific Value Judgments & Self Correcting Science

67:50 Choice in Metrics and Research Focus

70:30 Concluding Ideas

Watch it on...  YouTube: https://youtu.be/SzQn9SPVhRU Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-s02-e04-intro-to-abductive-reasoning-for-data-scientists/

The third and final of our (planned) short tutorials on key modes of critical reasoning. Abduction is common called "inference to the best explanation"...so it's easy to see why this concept is important for data scientists. 

Huub Brouwer (Utrecht University) walks us through a brief tutorial on how even a world-famous infer-er can get this wrong and how data scientists can avoid the same mistake.

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

0:00 Intro 0:18 Example of Abduction in Action 4:55 Definition of Abduction 6:21 Applying Abductive Reasoning 8:35 Why is Abduction Not Deduction? 14:55 Abduction in Data Sciences 17:40 Conclusion

Watch it on...  YouTube: https://youtu.be/lNOUvOUE_KE Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-s02-e03-intro-to-inductive-reasoning-for-data-scientists/

New episodes of the Philosophy of Data Science Series will now be published on Mondays!

Today's episode is a short introduction to a fundamental concept. Definitely worth your time!

Inductive reasoning is the fundamental challenge to scientific rigor. Induction is baked into methods like K-fold cross validation or generalizing from a sample to a population. However, many statisticians and data scientists are unfamiliar with the term and its implications. Joseph Wu (Brown University) gets us up-to-speed with a 10-minute presentation on the fundamental role of induction in scientific reasoning.

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

Outline 0:00 Intro 0:18 Inductive vs Deductive Reasoning 2:35 Overview of Induction, Deduction, and Abduction 3:23 Types of Induction: Everyday Life vs Statistical Generalizations 5:25 Sample to Population Induction 6:48 Population to Individual Induction 9:35 The Problem of Induction 11:52 Induction: Fallible but Powerful

Watch it on...  YouTube: https://youtu.be/y93D-55wgX8 Podbean: 

Deductive reasoning pervades statistics and data science...but how far can it get us to the right conclusion from data? Elina Vessonen (Finnish Institute of Health) gives a great 20-minute presentation reviewing the role of deduction in scientific reasoning. Elina begins with a common statistical example and then covers common deductive fallacies and their role in science.

It's a short and gentle introduction to a fundamental concept. Definitely worth your time!

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation. 

Thank you for your time and support of the series! 

0:00 Intro 0:18 Deduction Example in Statistics 4:05 Deductive Reasoning: Basic Concepts 6:42 Deductive Reasoning in Science 11:00 Falsification 16:05 Deductive Reasoning: A Summary

Session 2: Essential Reasoning Skills for Data Science

Episode 1: Round Table on Essential Reasoning Skills for Data Science

Session 2 "Essential Reasoning Skills for Data Scientists" is kicking off with a roundtable discussion with Elina Vessonen (Finnish Institute for Health & Welfare), Joseph Wu (Brown University), and Huub Brouwer (Tilburg University & Utrecht University).

One of the major challenges in data science is that we use three different modes of critical reasoning (deduction, induction, and abduction) on a daily (or even hourly) basis. It's important to understand the strengths and weaknesses of each mode of reasoning so that we can apply them as appropriate. This round table will begin this conversation on the modes of reasoning and how it applies to & science and data science.

Watch it on...

YouTube: https://www.youtube.com/watch?v=5bOuy6VA8Hg

Podbean: https://podofasclepius.podbean.com/e/philosophy-of-data-science-s02-e01-round-table-on-essential-reasoning-skills-for-data-science/

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

0:00 Intro

0:10 Roundtable on Critical Reasoning Skills

2:50 Guest Introductions

5:48 Thesis: Data Science Use All Modes of Reasoning Daily

6:45 Taxonomy of Deduction, Induction, and Abduction

17:21 The Problem of Induction

32:18 The Problem of Induction Creeping into Deduction

36:45 Bayesian Applicability Indices and Signal Quality Indices

40:55 What is "The" Scientific Method?

44:08 What is Pseudo-Science?

47:35 Theory vs Data/Evidence

50:48 Final Remarks

Session 1: Scientific Reasoning for Practical Data Science

Episode 3: Communicating the Science in Data Science

One of the biggest challenges in data scientists is to communicate why your work matters. Kathy Ensor (ASA 2022 President and Rice University’s Noah Harding Professor of Statistics) covers how to distinguish yourself as a professional by communicating both your scientific and technical value. (Hint: The same scientific reasoning that helps you do good work in data science will also help you critically assess “how” and “what” to communicate.)

Watch it on...

YouTube: https://youtu.be/Vtasc0GGKDs

Podbean:

You can join our mail list at: https://www.podofasclepius.com/mail-list

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

Scientific reasoning plays an essential role in data science and statistics, both for developing new methods and applying our methods to real-world problems. In Session 1's titular episode, Andrew Gelman talks through the role of scientific thinking in his approach to data analysis. He also highlights the good ideas that have been generated by the wider statistical community. 

Watch it on... YouTube: https://youtu.be/R6mq5Esjzfw

Coming up next week: Communicating the Science in Data Science with Kathy Ensor (Rice University & 2022 ASA President)

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

You can join our mail list at: https://www.podofasclepius.com/mail-list

#datascience #statistics #machinelearning #ai #science #stem

Session 1: Scientific Reasoning for Practical Data Science

Episode 1: Critical Reasoning in Medical Machine Learning

Data science in medicine and healthcare requires not only algorithmic and statistical knowledge but also a strong appreciation of the clinical environment in which (i) the data is being collected and (ii) the algorithm will be used. I'll showcase a scenario where a machine learning system failed to perform a "simple" clinical task and how critical reasoning was used to resolve the problem.

Guest-host Kristin Morgan (University of Connecticut) joins us to lead the discussion in how this example is applicable to the broader field of biomedical data science.

This is...

Session 1: Scientific Reasoning for Practical Data Science

Episode 1: Critical Reasoning in Medical Machine Learning

Watch it on... YouTube: https://youtu.be/o5YmdoCiyug

Podbean:

Coming up next week: Applying Scientific Reasoning to Statistical Practice with Andrew Gelman (Columbia University)

We're always happy to hear your feedback and ideas - just post it in the YouTube comment section to start a conversation.

Thank you for your time and support of the series!

Session 1: Scientific Reasoning for Practical Data Science

Episode 0: Welcome to the Philosophy of Data Science Series!

This is our very first episode of "The Philosophy of Data Science" series on Pod of Asclepius!

We go over our plans for the series plus some thoughts on why data science is such a rich field for discussions on scientific reasoning. Your time is valuable and you deserve a good explanation of why the topics were chosen and how the series is structured to maximize learning.

Topic List

0:00 New intro jingle for the series!

0:10 Welcome to the Philosophy of Data Science Series!

1:07 Modes of reasoning

5:33 Session 1 Overview: Scientific Reasoning for Practical Data Science

10:15 Session 2 Overview: Essential Reasoning Skills for Data Science

11:32 Keynotes and Session 4

14:15 Future Sessions

Coming up next week: Critical Reasoning in Medical Machine Learning

Thank you for your time and support of the series! It only gets better from here! (Seriously, it really does only get better from here. We've got Andrew Gelman coming up, plus Cynthia Rudin, Mihaela van der Schaar...)

She give a high-level overview of issues surrounding Innovative Trial Design and Master Protocols. A great listen for anyone wanting to be introduced to the subject or (for those already familiar) interested in its growing breadth of applications.

#datascience #statistics #biopharm #pharma #FDA

0:00 Intro

1:25 RelationalAI

3:25 Advantages of Julia as a foundation

4:21 "Full stack" data science

5:38 Advantages of Julia in the tech stack

6:30 Technical requirements of RelationalAI

7:45 Advantages of Julia (cont.)

10:00 Data munging, preprocessing, and transparency

14:30 Advantages of Julia (cont.)

18:35 RelationalAI's Innovation

22:00 Data Analysis and taking computational efficiency for granted

23:38 Who are the users of RelationalAI?

25:45 What are "knowledge graphs"?

28:30 Knowledge graphs for AI and Software 2.0

32:43 Julia as "executable math"

34:10 "Multiple dispatch" in a nutshell

36:20 Julia in the scientific community

38:53 See Nathan and Molham again at JuliaCon 2020

You can find the Virtual Undergraduate Career Fair here: https://ww2.amstat.org/virtualcareerservice/

Stephanie will be returning in a few weeks to discuss her own research at the nexus of data science, genomics, and public health.

Mihaela van der Schaar will be following up in a subsequent episode on this subject.

Shane shows us some of the real-time data analytics platforms that track the health of dairy cows and honey bees.

Prof. Jeroen Bergmann and Daniel Mogefors from the Oxford Healthtech Labs break down the key aspects of needs-led innovation and how researchers at Oxford University are using it

A quick explanation of who is coming on and why Glen has organized the episodes this way.

In Part 1 of the episode Allison talked about her background, research interests and the research that earned her a win in the ASA Student Paper Competition in 2017.

This is Part 1 of a two-part episode in which Allison talks about her background, research interests and the research that earned her a win in the ASA Student Paper Competition in 2017.

In Part 2 she will treat the audience to a technical deep-dive into optimizing bespoke clinical models.

This part discusses the MDD Idea Exchange and Bayesian p-values.

The other two parts of this episodes cover:

Part 1: MDD Section Activities Part 2: Bayesian Methods and Digital Health Initiatives

This part discusses Bayesian Methods and Digital Health Initiatives.

The other two parts of this episodes cover:

Part 1: MDD Section Activities Part 3: The MDD Idea Exchange and Bayesian p-values

This part discusses the MDD Section Activities. 

The other two parts of this episodes will cover:

Part 2: Bayesian Methods and Digital Health Initiatives Part 3: The MDD Idea Exchange and Bayesian p-values

In this part, Glen describes a nifty statistical approach to vital-sign artefact detection.

In the first part of this 2 part episode, Glen talks about using Gaussian Processes for identifying the deteriorating patient.

In this part Glen talks about Gaussian Processes for Identifying the Deteriorating Patient. 

The second part of this 2 part episode will describe a nifty statistical approach to vital-sign artefact detection.