Software
weathermetrics
The heat index, commonly called the "feels like" temperature, is a
popular metric of heat exposure in environmental health research.
However, the algorithm used to calculate heat index varies across
environmental studies, and the algorithm used by the US National Weather
Service is rarely used.
With permission from members of the US National Weather Service, I converted their heat index algorithm into an R function. This function allows simultaneous calculation of heat index for a series of weather observations. Inputs can be temperature in either Fahrenheit or Celsius and either dew point temperature or relative humidity.
The "weathermetrics" package also includes functions to convert temperature observations between Fahrenheit and Celsius, to determine dew point temperature from air temperature and relative humidity, and to determine relative humidity from air temperature and dew point temperature.
This package is freely available through the Comprehensive R Archive Network (CRAN) and can be installed on any R installation using the "install.packages" function.
bespoke
Web applications offer several advantages over traditional homework
or practice problem sets. Students receive immediate feedback and can
continue trying difficult problems until the material is mastered.
Additionally, computers can easily randomize numbers and data for a single
question template to generate multiple examples of the same type of
question.
Despite these advantages, development of web applications from scratch requires programming knowledge many academics lack. R, an open source programming language, is growing in popularity in many academic fields, and because of R's popularity in academic research, a growing number of professors are comfortable coding in R.
Roger Peng and I are developing software that allows professors or teaching assistants to create templates for practice problems exclusively using R. Once templates are created, a function in the package randomizes each template to create multiple versions of each question and then structures the questions within a framework of PHP and html code to create a web application.
We have used this software to create custom web applications for two Master's level courses at Johns Hopkins Bloomberg School of Public Health. After further refinement, we plan to post this as open source software on the Comprehensive R Archive Network (CRAN) for others to use.
Interview with John McGready about bespoke
On October 11, 2012, John McGready interviewed me for his "Famous Statisticians" video series. We talked about my "bespoke" software package, which we used to create a web application with practice problems for his Masters' Biostatistics course.