Mining Twitter for Airline Consumer Sentiment
by Jeffrey Breen

Airlines, Consumers, and Twitter

Anyone who travels regularly recognizes that airlines struggle to deliver a consistent, positive customer experience. Through extensive interview and survey work, the American Customer Satisfaction Index ( quantifies this impression. As a group, airlines falls at the bottom of their industry rankings, below the Post Office and insurance companies:

Read more

Time Series Analysis and Order Prediction with R
by Prof. András Pataricza, Prof. Gábor Horváth, Imre Kocsis, Kristóf Gáti. Department of Measurement and Information Systems, Budapest University of Technology and Economics, Budapest, Hungary

Predictive analytics in manufacturing uses models extracted from historical order and production related records to support the optimization of manufacturing processes. For instance, forecasting the overall amount of incoming orders is crucial for a) properly scheduling and sizing raw material and part orders and b) configuring processes as e.g. final assembly or testing. Loosely based on a specific subproblem in one of our ongoing R & D projects, this howto showcases the process of building an order amount predictor in R.

Read more

by Matthew J. Flynn, Ph.D

R in business

My application of R in business is a method to aid two groups of R users at our company.

Read more

Using R to Estimate the Pythaogrean Won-Loss Formula in Hockey
by Kevin Dayaratna

I have a paper under review at a journal where I (along with my co-author) provided theoretical verification for applying the Pythagorean Won-Loss formula to hockey. Originally devised for baseball, the Pythagorean Won-Loss formula estimates the percentage of games a team should have won at a particular point in a season. For decades, this formula had no mathematical justification from first principles. In 2006, the second author provided a statistical derivation by making some heuristic assumptions about the distributions of runs scored and allowed by baseball teams.

Read more

A Direct Marketing In-flight Forecasting System
by Shannon Terry & Ben Ogorek, Nationwide Insurance


R's powerful built-in functions and ease of web-based deployment led to its role in the In-flight Forecasting System, a browser-based utility designed to forecast the total incremental benefit of a marketing tactic when only a fraction of the marketing responses have been observed. This article shows how R can be used to estimate this total incremental benefit early, as well as how R can be run in an interactive, business-friendly format over the web to deliver this methodology to the business.

Read more

Predict the duration of your clinical trial
by simon66217

Researchers overpromise and undeliver on patient accrual, the time frame in which they plan to obtain the proposed sample sizes for their research studies. It is not uncommon for a researcher to promise to get 100 patients within a year, but then struggle to get even a dozen patients after two years. Slow patient accrual leads to delays in completion of the study or sample size shortfalls or both.

Read more