William M. Briggs

Download a PDF copy of my February 2010 Quirk’s article: Telling the future from the past: predictive versus classical statistics. If you want to read the article online at Quirk’s, registration is required, but free.

 

The old ways of doing statistics gave too much certainty—they produced overconfidence. They told you what you already know. This led to false comfort. Worse, they guaranteed costly mistakes.

Predictive analytics tell you what you do not already know. It gives precise measures of uncertainty of real-world matters of interest. It allows you to make real decisions about real problems. And it does it directly.

Most statisticians still use the old ways because those old ways are still what are taught. The inertia in the system is tremendous. This means the new ways are known to only a few.

If you have to make real money decisions, give me a call or send me an email and I’ll be happy to explain how the new ways of doing statistics can help. See below about Lectures.

Statistician to the Stars!

It’s time for you to join the ranks of luminaries that have hired William M. Briggs. From the smallest to the largest job imaginable, no person can enhance your celebrity more than he can.

I am one of the very few practitioners outside of a university using the most modern methods.

Email matt@wmbriggs.com or call today to find out how I can help you.

Click here for a list of statistics articles.

Modern statistics

Still using p-values? Estimating coefficients from regression models? The old ways of statistics had their day, but it’s time to move up to the modern way.

Predictive analytics shifts thinking away from the data you have—about which you know everything—to data you have not seen. That data is the only true unknown.

Lectures Available

I can lecture about modern statistics to executives, managers, those who just need to understand statistical results, or to other statisticians who want to catch up on predictive techniques. See my Lectures page for more details.

What? You don’t have web access for your clients?

Are you a marketing, insurance, credit or other company that relies on creating analyses for clients and you don’t yet have a web portal for these customers? Why not?

Talk to me about how to build automatic analytic web-engines.

Review of your statistical services

Data collection should be fully integrated with data analysis. It should move from collection to database to analysis automatically. And that analysis should be using modern methods. Ask me about a complete review of your analytic processes.

Sample of expertise

Hard Core Models	Predictive Models	Mapping
Regression	PLS	Bayesian networks
Choice models	MARS (splines)	Affinity analysis
Text mining	Support vector	Correspondence analysis
Max Diff	Neural networks	Collaborative filtering
Hierarchical/mixed	GAM	SWOT analysis
Fraud detection	Time series/GARCH	Cluster analysis
Tree models	Forecasting	Multi-domain analysis
Design	Choice Models	Coding
Experimental	Discrete	R
On-line ad campaigns	Pairwise	SAS
Large datasets	Multinomial	PHP
		Many others

Contacting me

When is the best time to bring in a statistician? Before you begin collecting any data. I’ll be happy to talk to you, free of charge for the first consultation, to explore the best way you can go about answering the questions you have. My emphasis is on understanding, on reducing complexity of results, and on finding simple, clear pictures to show your audience what you want to say.

I have been involved in just about every kind of statistical analyses under the sun (some on and beyond the sun, too: see my resume). Please feel free to contact me about any questions you have about uncertainty, probability, the best places to get steak and kidney pie in New York City, or even statistics at matt@wmbriggs.com.

How much?

Rates by the hour, by the job, or by retainer are available. I’m not a stickler with the clock: you’ll get more than your money’s worth.

I am very willing to work with small and growing companies. I have been involved in establishing many interesting, non-traditional partnerships.

This means: the stranger your job, the more I’ll like it.

More Services

Custom Analysis	Any and all kinds of data analysis, including—especially including—advice on how to and what data to collect. I’ll help put your data in the best light possible and to fully understand the uncertainty inherent in the situtation. I offer unique, original, clever analyses that you won’t find in too many places.
Data mining	Advice, custom code, automatic code, web-based data analysis systems, graphical analysis and data display, machine learning, data mining, and so on. See my resume for past experience in this area.
Survey design & analysis	The best time to contact me is before you begin. I’ll help avoid the common mistakes so that you do not feel despair after the data is collected.
Weather and climate insurance	See my white paper on this topic.
Research design & analysis	This includes the standard sample size and “power” analyses, but also, and much more importantly the design of experiments. I cannot tell you how many times people have come to me after they have run their experiment, only to discover that the questions they wanted answered cannot be answered with the data they have collected.
Forecasting and forecast analysis	Attention stock pickers, money managers, business forecasters. Forecast analysis? Yes, sir. If you are, right now, making use of predictions, forecasts, prognostications, etc., you owe it to yourself to discover how good these models are. Could you be making more money by modifying your forecast?
Business intelligence	The whole gamut of marketing research: Q-sort, (I’ll try to talk you out of) factor analysis, custom code, custom design of computer systems, and on and on.
Bayesian modeling	This is it. The most modern, sophisticated form of data analysis available. If you’ve been keeping track of my blog, you’ll see the hundreds of reasons why this is not only the better kind of analysis, but the proper one, too. Feel free to call or write and ask questions.
Training	Click over to my Lectures page to see the broad range of training, seminars, lectures, and tutoring I offer.
Market research	Market, customer segmentation analysis, Maxdiff, Random forest and decision trees, PLS, 3-D graphics and modeling, sentient analysis, cluster and factor analysis, market response, forecasting.

Privacy

All data is securely stored and no results are ever shared with anybody except the clients. Client names, unless permission is given, are also kept confidential.