040: William MacMillan (Prattle)

William MacMillan on Innovation City

“One of the challenges people face with information and data really is – you kind of don’t know how important things are… and it becomes really difficult to decide where to allocate your attention.” — William MacMillan

Welcome to Innovation City—powered by Venture Cafe—where Tyler Kelley and Michael Johnson, Co-Founders of SLAM! Agency, interview innovators, creators, and disruptors to discover how business is changing in the modern world.

Created and produced by SLAM! Agency in conjunction with Venture Cafe St. Louis and Venture Cafe Miami, Innovation City gives you an inside look at how rapidly business and culture are changing thanks to increasing diversity and inclusion, heightened creativity, and a stronger and better-connected business community. Venture Cafe is the largest combined gathering of entrepreneurs and innovators anywhere in the world. Events are held every Thursday in St. Louis, Miami, and other leading innovation cities around the globe.

Today’s guest is William MacMillan, Ph.D., CTO & Co-Founder of Prattle. An experienced corporate data scientist, fraud analyst, and former professor at Washington University and the University of Michigan, Bill has performed extensive research on government finance and is the co-author of ‘How the Fed Moves Markets.’ Bill is the author of original software that estimates trade flows, exchange rates, government policy, and policy regimes. Prattle quantifies market-moving language, algorithmically generating data and reports on communications from central banks and publicly-traded companies. Bill joins the Innovation City Team to discuss the creation of the field of data science, Prattle’s growing role in bringing attention to under-researched central banks in countries all over the world, and the value of statistical reasoning. Find Prattle on Twitter @prattledata.

They discuss:

  • Bill’s standing in the top rank of practitioners in the field of data science
  • The history and development of data science, and the breadth of types of projects that can be performed with varying amounts of data
  • Data scientist: being a statistician, but using a slightly different method
  • Bill’s Roomba-eque path to becoming a data scientist
    • Took (maybe) one math class as an undergrad
    • Studying statistics as a graduate student
  • By the time he finished his Ph.D. Bill had a strong command of statistics, but data science as a field was still in its infancy
  • Statistical methods that serve as the tools of data science; neural nets, lassos, econometrics, and other models
  • The tremendous amounts of data that need to be collected for effective research, and why the private sector has outpaced the educational sector in this regard
  • Applied data science vs. what can be learned in school
  • His time as a genomicist at Monsanto
    • Asking coworkers lots of questions because he had no biology training
  • Performing quality assurance on genomics testing
  • How data science models carry over across different use cases: patterns of speech analysis, image analysis, election fraud, and more, all from the same computational model
  • Information deficits and perfect data
  • The unique characteristics of language that moves markets
  • Other text analytics vendors use one model across varying instances
  • Prattle takes a top-down approach to solving language consumption problems for market actors
  • Investor relations; ‘Does this data matter?’ …Prattle can answer that for you
  • Prattle’s clients: working with investment banks, hedge funds, mutual funds, futures funds, commodities firms
  • Where A.I. and Data Science are headed
  • Robo-advisors in the financial space
  • High-net worth individuals need whole businesses devoted to managing their money
  • A.I. posing a threat to white collar jobs
  • Bots all the way down: how many duplicate articles are built from a base piece of content
  • The value of statistical reasoning: it improves outcomes in every field
  • The isolation of academic work
  • Bill’s background as a construction worker, and how highly he values teamwork
  • Bill’s goal of improving the state of the American worker
  • Prattle’s research automation helps unwatched companies get coverage, find investors, and create jobs
  • Central Banks in other countries have asked Prattle to cover them in order to boost foreign investments
  • Prattle’s different products each look at a minimum of several hundred thousand documents per day
  • Prattle as Moneyball for financial markets
  • Human minds vs. analytical A.I. minds; finding edges that human beings can’t exploit
  • The increasing quantification of data in sports; teams performing better