no name

Shared posts

21 Jul 18:22

The Statistics Wars and Their Casualties Workshop-Now Online

by Mayo

The Statistics Wars
and Their Casualties 

22-23 September 2022
15:00-18:00 pm London Time*

ONLINE 

To register for the workshop, please fill out the registration form here.

*These will be sessions 1 & 2, there will be two more
The future online sessions (3 & 4)  at 15:00-18:00 pm London Time on December 1 & 8.

Yoav Benjamini (Tel Aviv University), Alexander Bird (University of Cambridge), Mark Burgman (Imperial College London),  Daniele Fanelli (London School of Economics and Political Science), Roman Frigg (London School of Economics and Political Science),
Stephan Guttinger (University of Exeter), David Hand (Imperial College London), Margherita Harris (London School of Economics and Political Science), Christian Hennig (University of Bologna), Daniël Lakens (Eindhoven University of Technology), Deborah Mayo (Virginia Tech), Richard Morey (Cardiff University), Stephen Senn (Edinburgh, Scotland), Jon Williamson (University of Kent)

While the field of statistics has a long history of passionate foundational controversy the last decade has, in many ways, been the most dramatic. Misuses of statistics, biasing selection effects, and high powered methods of Big-Data analysis, have helped to make it easy to find impressive-looking but spurious, results that fail to replicate. As the crisis of replication has spread beyond psychology and social sciences to biomedicine, genomics and other fields, people are getting serious about reforms.  Many are welcome (preregistration, transparency about data, eschewing mechanical uses of statistics); some are quite radical. The experts do not agree on how to restore scientific integrity, and these disagreements reflect philosophical battles–old and new– about the nature of inductive-statistical inference and the roles of probability in statistical inference and modeling. These philosophical issues simmer below the surface in competing views about the causes of problems and potential remedies. If statistical consumers are unaware of assumptions behind rival evidence-policy reforms, they cannot scrutinize the consequences that affect them (in personalized medicine, psychology, law, and so on). Critically reflecting on proposed reforms and changing standards requires insights from statisticians, philosophers of science, psychologists, journal editors, economists and practitioners from across the natural and social sciences. This workshop will bring together these interdisciplinary insights–from speakers as well as attendees.

Sponsors/Affiliations:

The Foundation for the Study of Experimental Reasoning, Reliability, and the Objectivity and Rationality of Science (E.R.R.O.R.S.); Centre for Philosophy of Natural and Social Science (CPNSS), London School of Economics; Virginia Tech Department of Philosophy

Organizers: D. Mayo, R. Frigg and M. Harris
Logistician (chief logistics and contact person): Jean Miller
Executive Planning Committee: Y. Benjamini, D. Hand, D. Lakens, S. Senn

To register for the workshop,
please fill out the registration form here. 

View attached file (e6d70454df5771d987169e38da9c9ac7?s=96&d=https%3A%2F%2F2.gravatar.com%2Favatar%2Fad516503a11cd5ca435acc9bb6523536%3Fs%3D96&r=PG, unknown)
14 Nov 17:56

PB Python Article Roadmap

by Chris Moffitt

Introduction

September 17th is Practical Business Python’s anniversary. Last year, I reflected on 5 years of growth. This year, I wanted to take a step back and develop a guide to guide readers through the content on PB Python.

As of this writing, I have 84 articles on the site. They vary from fairly complex and lengthy to quick summaries. When I wrote them, I did it based on my interests at the time and without much thought on progression. Now that I have a decent volume of articles, I want to organize them in a more meaningful way.

My ultimate goal for this site is that I want it to be a resource to help people use Python to automate away many of the repetitive tasks they do on a daily basis with tools like Excel. A secondary goal for is to cover more advanced Python topics that are difficult to do in Excel.

One reason for developing this guide is that at least 90% of my traffic comes from organic search. These users come to the site, read an article and move on. I hope those that find this guide will stay a little longer and find other relevant content and use this as a resource to navigate the Python ecosystem.

Secondly, this guide will be useful to help me understand gaps in the content and keep a mental framework for continuing to develop content. My intent is to update the sections below as I add new content. I will also maintain a link at the top of the Archives page to point people in the right direction.

Getting Started with Python

Before you begin your Python journey, here are a couple of articles that are helpful for getting everything set up on your system:

Just as importantly, you need to think about how long this journey will take and what you may need to do to spread the knowledge within your organization:

Case studies are also a great way to understand how the Python ecosytem can be used to solve problems:

Pandas Fundamentals

Pandas is a rich library with a lot of functionality. If you are new to pandas, this is the order I would recommend reading the articles:

Basic pandas concepts:

Grouping and summarizing data:

Data input and output:

Data Science Topics

Data science is a broad category and many articles are related to various data science topics. These articles are more focused on specific data science tasks:

Data Visualization

Python’s data visualization landscape is complex and it can be difficult to determine the best tool to use. Here are some posts about the visualization landscape:

Articles on some specific libraries.

Matplotlib:

Altair:

Bokeh:

Plotly:

Seaborn:

Site Updates

This section contains ad-hoc posts about the site and the technology behind it.

04 Jun 20:07

Blockchains and the Opportunity of the Commons

by Alex Tabarrok

Tyler asks which goods and services are most likely to be bought and sold on a blockchain that is paid for with token issuance and appreciation?

  1. The services with high mark-ups? Low mark-ups?
  2. Big consumer bases?
  3. Well informed and well coordinated consumer bases?
  4. “Influencer” consumer bases, in the Gladwellian sense?
  5. “Trivial” consumer bases, that you don’t mind risking?
  6. Some other properties?

I will go with 6. Blockchains and tokenization are a way to incentivize the creation of a commons. A commons is an unowned place, platform, or protocol that helps people to meet, communicate and transact. Commons underlying modern life include TCP/IP, SMTP, HTTP, GPS and the English language. We don’t see these commons clearly because they are free, ubiquitous and, like air, taken for granted. What we do see are platforms like Airbnb, Uber and the NYSE and places to meet and communicate like OkCupid, Twitter, Facebook and YouTube. What blockchain and tokenization offer is the possibility of creating commons to replace all of these services and much more.

As the examples of AirBnb, Facebook and YouTube indicate, it’s possible for private firms to create platforms that serve the same purposes as a commons but these platforms are not a commons since they are privately owned. Private ownership is great but not without tradeoffs. Bill Gates hinted at one problem when he defined a platform:

A platform is when the economic value of everybody that uses it, exceeds the value of the company that creates it.

The platform dilemma is that a company that controls a platform wants to maximize the company’s value rather than the economic value of everybody that uses it. Company value and social value are correlated but they are not the same. There are three problems. First, the company will want to grab up as large a share of the social value as possible. That’s ok for efficiency but not ideal for platform users who, because of network effects and coordination issues, may find that they need to use the platform even though it leaves them with only a small surplus. Second, the company may take actions that increase its value but reduce social value. On some margins, for example, Facebook and YouTube profit from advertising that reduces social value. The third problem is that in creating a platform where many people meet and transact, a small number of companies come to control and access more data than may be ideal. Big centralized data is worrying for libertarian reasons but also because big, centralized data is a honeypot for bad actors and hence insecure.

The first set of internet commons like TCP/IP and HTTP were created by government and independent researchers. The unique use-case of blockchains is that blockchains can be used to incentivize the creation of unowned platforms, i.e. commons. The creator of a blockchain need not control the blockchain and indeed can credibly commit not to control it. Thus, the creator of a blockchain can commit to never taking actions to maximize profit at the expense of social value and it can commit to never taking actions to redistribute more of the social value to itself. The blockchain creator, however, can be rewarded through token issuance. Moreover, since the value of the token and the social value of the blockchain are positively correlated the blockchain creator has strong incentives to create a commons that maximizes social value.

To give an example, LBRY–one of the blockchain firms that I advise–is a kind of YouTube on the blockchain. The protocol that LBRY has created is unowned. LBRY’s incentives are to create something that will maximize the value of both content creators and content consumers. The social value created could well exceed that of any owned platform and if LBRY earns a small share of this social value they will be well compensated. Token issuance and appreciation is what incentivizes the creation of the commons.

Creating a commons on the blockchain isn’t easy, however. Decentralized institutions are much more difficult to design than centralized institutions. Decentralized databases are a big advance but making them work at scale-size and speed is a challenge. Precisely because the blockchain is unowned the designers have to get much more correct, right out of the gate. Changing a commons on the fly, forking, is costly, disruptive and not always possible. All of this explains why in the history of the world almost all decentralized institutions, such as markets and language, were not designed but arose through evolutionary forces. Hayek called decentralized institutions spontaneous orders because he implicitly assumed that all such decentralized institutions were spontaneous, i.e. unplanned. Only in very recent years have economists and computer scientists developed the understanding and tools that are necessary to design decentralized orders–orders that are planned but not controlled. Today smart contracts on blockchains like Ethereum have the potential to create a sophisticated set of global common resources that will form the foundation for much of the economic and social structure of this century–this is the opportunity of the blockchain commons.

The post Blockchains and the Opportunity of the Commons appeared first on Marginal REVOLUTION.

23 Mar 20:08

The rise and fall of cognitive skills

by Alessandro Cerboni

Neuroscientists find that different parts of the brain work best at different ages.

Scientists have long known that our ability to think quickly and recall information, also known as fluid intelligence, peaks around age 20 and then begins a slow decline. However, more recent findings, including a new study from neuroscientists at MIT and Massachusetts General Hospital (MGH), suggest that the real picture is much more complex.

The study, which appears in the journal Psychological Science, finds that different components of fluid intelligence peak at different ages, some as late as age 40.

“At any given age, you’re getting better at some things, you’re getting worse at some other things, and you’re at a plateau at some other things. There’s probably not one age at which you’re peak on most things, much less all of them,” says Joshua Hartshorne, a postdoc in MIT’s Department of Brain and Cognitive Sciences and one of the paper’s authors.

“It paints a different picture of the way we change over the lifespan than psychology and neuroscience have traditionally painted,” adds Laura Germine, a postdoc in psychiatric and neurodevelopmental genetics at MGH and the paper’s other author.



23 Jan 16:10

The North Carolina unemployment insurance experiment may be looking up

by Tyler Cowen

The benefits have been stopped, and there has been much recent debate over how well this is working to stimulate reemployment.  This new study is from Kurt Mitman, who is a doctoral candidate at U. Penn and an NBER research associate, here is his summary:

1. Evidence from the establishment survey confirms a substantial increase in employment in North Carolina following the unemployment insurance reform.

2. The increase in payroll employment reported by the sample of North Carolina employers is smaller than the increase in employment reported by workers in the household survey.

3. The increase in employment [is] driven by the private service sector.

4. A comparison of the growth in employment between North Carolina and the adjacent states in Figure 5 reveals a similar growth in the post-reform period between the two Carolinas, which is much faster growth than in Virginia.

5. Results in Table 3 reveal a mild tendency toward higher weekly hours post reform and little change in wages and earnings.

The full piece is here (pdf).  This seems to me our best understanding of the admittedly limited data to date.