Blog

rss

Posted on Aug 14, 2019 By: Peter Bruce
It is hard to imagine a more traditional, sole-proprietor business than the custom (bespoke) tailor.  The advent of the web, ecommerce and the ability to submit measurements online was the first phase of consolidating and transforming custom tailoring.  The second phase, made possible by machine learning, is custom styling (not just manufacturing to measure). Stitchfix, the San Francisco startup founded by Katrina Lake and Erin Flynn, helps customers craft their own personal style by working w...
Posted on Aug 14, 2019 By: Peter Bruce
I’ve been told that, in conversation, I jump in and finish other people’s sentences for them.  Now there’s an app for that: Open AI Lab, founded by Elon Musk, has released GPT-2, its program that, given a prompt, will write (mostly) intelligible content.  Open AI Lab, which Musk recently left, has as its mission “to ensure that artificial general intelligence (AGI) … benefits all of humanity.” GPT-2 is a weaker version of the full program that Open AI has developed, but kept lar...
Posted on Aug 14, 2019 By: Peter Bruce
Hans Rosling popularized animated visualizations of data that dramatically revealed human progress in health and welfare that people often lose sight of.  Now check out Cedric Scherer’s step-by-step ggplot tutorial on how to turn a somewhat pedestrian boxplot display into a visualization that incorporates many levels of information.    
Posted on Aug 14, 2019 By: Peter Bruce
Chris Jerde "I’ve been using your curriculum to refresh and build upon what I offer to my clients. It’s been really helpful so far." Chris Jerde worked for 10+ years with the consulting arm of a large architectural firm. He has long been interested in the analytics side of the business - measuring what users need in terms of building performance and aligning that with what is built. Last year, he ventured out on his own, forming Jerde Analytics to offer these services to other clients. In ...
Posted on Aug 14, 2019 By: Peter Bruce
INFORMS Analytics Body of Knowledge This 370-page edited volume is a compendium of methods that are considered important for organizations deploying analytics, generally, and for the INFORMS Certified Analytics Professional (CAP) program in particular.  INFORMS is the US professional society devoted to operations research and analytics, with analytics being a recent addition to its portfolio. CAP is an exam-based certification process covering a variety of methods, including: Predicti...
Posted on Aug 14, 2019 By: Peter Bruce
How long do you think yellow traffic signals should last?  Mats Jarlstrom had definite opinions and shared them via email, noting that he was an engineer.  For his trouble, he was fined $500 by the Oregon State Board of Examiners for Engineering and Land Surveying for “practicing engineering without a license.” Should statisticians, analytics professionals and data scientists be professionally certified or accredited?   There are a variety of certifications available in the field of s...
Posted on Aug 12, 2019 By: Peter Bruce
Bandit Algorithms for Website Optimization, by John Myles White A classic statistical experimental design comparing treatments (two treatments, treatment versus control, multiple treatments) specifies a sample size, collection of data, then a decision, typically based on hypothesis-testing:  the winning treatment must attain a level of statistical significance, otherwise you go with the default “null hypothesis.”   This protocol is much too ponderous and slow for the world of web-...
Posted on Aug 09, 2019 By: Peter Bruce
A friend tosses two coins, and you ask “Is one of them a heads?”  The friend replies “Yes.” What is the probability that the other is a heads? ANSWER:   One-third.  There are four ways the coins could have landed originally: HH:  0.25 probability HT   0.25 probability TH   0.25 probability TT   0.25 probability (The probability that the first coin lands H is 0.5, and the probability that the second coin lands heads is 0.5, so the probability that each lands heads is...
Posted on Aug 09, 2019 By: Peter Bruce
A casual user of machine learning methods like CART or naive Bayes is accustomed to evaluating a model by measuring how well it predicts new data.  When examining the output of statistical models, they are often flummoxed by the profusion of assessment metrics. Typical multiple linear regression output will contain, in addition to a distribution of errors (residuals) and RMSE (root mean squared error), such values as R-squared, adjusted R-squared, t-statistics, F-statistics, P-values, degrees o...
Posted on Aug 09, 2019 By: Peter Bruce
I was visiting New York a couple of weeks ago, transferring from Amtrak to the PATH trains at Newark.  PATH takes you to Wall Street - the #1 financial center in the world - and yet the process of paying for my $2.75 PATH ticket was excruciating.  When I arrived at Newark, my colleague, who had arrived 30 minutes earlier on a different train, was still in line at a PATH farecard machine. Credit card transactions were taking minutes before being mysteriously denied.  The most efficient means o...
← Older post