Richard M. Heiberger Professor Emeritus of Statistics Temple University January 2016

Transcription

1 Richard M. Heiberger Professor Emeritus of Statistics Temple University January 2016 Graphical Display of Data, Data Analysis, Data Visualization, Design of Experiments, Design of Software Systems Books Computation for the Analysis of Designed Experiments, Wiley Statistical Analysis and Data Display (with Burt Holland), Springer 2015 (second edition) and 2004 (first edition). R through Excel (with Erich Neuwirth), Springer R through Excel has been translated into Japanese. Software While on research leave at Bell Labs ( ), I participated in the design of the lm and aov functions in S and R. I maintain two R packages that are distributed on CRAN HH: software for my book Statistical Analysis and Data Display RcmdrPlugin.HH: interactive access to the HH package I am a member of the core team that designed and maintains ESS (Emacs Speaks Statistics), a powerful editing system and user interface to R and other software. I wrote and maintain the RthroughExcelWorkbooksInstaller, software that provides examples for the book R Through Excel. 1

2 Research Contracts and Consulting GlaxoSmithKline, Philadelphia, PA ( , 2009). Consultant on Statistics and Computation AT&T Bell Laboratories, Murray Hill, NJ ( ) Consultant. Participant in design of the lm and aov functions in S and R. Subcontract with NORC for US Census Bureau project on Data Visualization Research. September 2012 January Subcontract with ICF International for US Department of Environmental Protection project on Exploration of Data Visualization. May 2012 July Heiberger, R. (Principal), User friendly interface for supply modeling tools, Sponsored by Glaxo- SmithKline ( ). CCPDS Tumor Registry Casefinding Study. Subcontract with Fred Hutchinson Cancer Research Center, National Cancer Institute (1980). Robustness of Algorithms for Regression and Analysis of Variance. National Science Foundation ( ) Evaluation of Computer Programs for Regression, Analysis of Variance, and Related Topics. National Science Foundation ( ). Short Courses UseR! 2012 R Users Conference, Nashville, TN (June, 2012) Tutorial on ESS, Short Course. Joint Statistical Meetings Toronto 2004, Minneapolis 2005, and other venues. Statistical Analysis and Data Display, Short Course. American Statistical Association Conference on Statistical Practice 2014: Structured Sets of Graphs. Tampa, Florida. 2015: Design of Not-Simple Graphs. Half-Day Course. New Orleans, Louisiana. 2016: Design of Multi-Panel Graphs. San Diego, California. Deming Conference on Applied Statistics, Atlantic City American Society for Quality Control and American Statistical Association 1990: Computation for the Analysis of Designed Experiments 2005: Statistical Analysis and Data Display 2016: Statistical Analysis and Data Display, Second Edition Recent Presentations Richard M. Heiberger, July 2015 Computational Precision and Floating-Point Arithmetic. UserR! 2015, Aalborg, Denmark. Richard M. Heiberger, July 2013 Design of graphics with lattice and mosaic. UseR! 2013, Albacete, Spain. Richard M. Heiberger, Edward J. Mulrow, Naomi B. Robbins, August 2013, Visualizing Census Tables, Joint Statistics Meetings, Montreal, Canada. 2

3 1st Annual FDA/DIA Statistics Forum. FDA/DIA Federal Drug Agency/ Drug Information Association. Heiberger, Richard M. (2007). Graphical Approaches to the Analysis of Safety Data from Clinical Trials. Bethesda, MD, March 7, 2007 Education Ph.D., Harvard University (1972) Statistics M.A., Harvard University (1968) Statistics A.B., Oberlin College (1967) Mathematics A.B., The Central High School of Philadelphia (1963) Academic Positions Temple University, Department of Statistics Professor Emeritus 2012 Professor Associate Professor Honors Faculty Mentor for the Future of Instructional Technology, Office of the Provost, Courses Statistics 8121 Statistical Computing I Statistics 9190 Statistical Computing II The overall description for the two courses, with about half the material in each semester, is: Use of computers in the solution of statistical problems. Topics include: floating point architecture, random number generation, design of statistical software, computational linear algebra, numerical integration, optimization methods, spreadsheets, interprocess communication (including ESS, RExcel, operating systems), the construction of packages of R functions, the FAQ files in R, data quality, language design (including parsing and text utilities), editors, graphics, color perception, simulation, EM, MCMC. Term projects from the last few times I have offered the courses have led to presentations at international conferences, articles submitted to refereed journals, and successful PhD dissertation proposals. What the course is NOT: This is NOT a course on statistical methods. The prerequisite course ( ) shows how to use statistical software for the analysis of data. This course (8121) discusses how the software systems are themselves constructed and how they work. Statistics 8107 Design of Experiments I Principles of experimental designs, completely randomized designs, multiple comparisons, randomized block design, latin square design, missing value problems, analysis of covariance, and factorial experiments. 3

4 Statistics 8003 Statistical Methods I A prior course in applied statistics, essentially the standard introductory course that is part of most undergraduate programs (Stat 2022 at Temple), is a prerequisite for Stat My basic working mathematics assumption is a knowledge of algebra, elementary differential calculus (use of derivatives to find maxima and minima), and elementary matrix operations (including eigenvalues and eigenvectors). I expect students enrolled to have some aptitude in mathematical areas, and if deficient in background, to be willing to take extra time in the beginning of Stat 8003 to acquire additional background. Introduction to frequently used methods. Includes probability, estimation, tests of hypothesis, survey sampling, linear regression data analysis, statistical computer packages. Statistics 8004 Statistical Methods II Introduction to multiple regression, analysis of variance, design of experiments, robust techniques, analysis of covariance, nonparametric analysis, and multivariate analysis. Statistical packages such as R and SAS are used. Prerequisite: Stat Statistics 8102 Statistical Methods III Linear and non-linear mixed effects modelling, generalized linear models, missing values, survival analysis, multivariate analysis, Bayesian statistics, data mining, spatial techniques, neural networks. Prerequisite: Stat Recent PhD Students Keith Halbert (2014), Estimation of Probability of Failure for Damage-Tolerant Aerospace Structures. Kathleen Campbell (2014), Extension of Kendalls τ Using Rank-Adapted SVD to Identify Correlation and Factions Among Rankers and Equivalence Classes Among Ranked Elements. Min Fu (2014), A Resampling Based Approach in Dose-Response Drug Research. University of Pennsylvania, Department of Biostatistics Associate Professor Assistant Professor University of Pennsylvania, Department of Statistics Assistant Professor Iowa State University, Department of Statistics Assistant Professor

5 Associations and Honors American Statistical Association Fellow of the American Statistical Association Chair of the Section on Statistical Computing for 2011 Member of the Special Task Force on Student Posters of the Statistical Graphics Section 2012 President of the Philadelphia Chapter Committee on Certification of Statisticians Recent Papers Design of Diverging Stacked Bar Charts for Likert Scales and Other Applications, with Naomi Robbins (2014), Journal of Statistical Software The software is included in the HH package in R. The graphical technique for displaying Adverse Event data that we published during my research leave at GlaxoSmithKline in , and also available in the HH package, is now standard in the pharmaceutical industry. (Ohad Amit, Richard M. Heiberger, and Peter W. Lane. (2008) Graphical Approaches to the Analysis of Safety Data from Clinical Trials. Pharmaceutical Statistics, 7, 1, Richard M. Heiberger and Burt Holland (2006), Mean Mean Multiple Comparison Displays for Families of Linear Contrasts, Journal of Computational and Graphical Statistics 15, 4, The software is included in the HH package in R. Much of my current work is based on the RExcel interface between R, the best statistical computing language, and Excel, the most prevalent spreadsheet software system. Placing R inside the Excel user interface brings high-quality, state-of-the-art statistical software to the end user. Through my work on RExcel I have been designing software and lesson plans for improving statistical education. I use this material in all courses I teach. See my book with Erich Neuwirth R through Excel listed above. Heiberger, R., Robbins, N. B. (2011). Alternative to Charles Blow s Figure in Newt s War on Poor Children. Administration In I was Acting Associate Vice Provost at Temple University, with primary responsibility ( ) for installing the DARS (Degree Audit Reporting System) into the University s computers. The DARS system matches students courses to the requirements for their declared major and minor. I was a member of the Statistics Graduate Committee and chaired it I was a member of the Statistics Undergraduate Committee and chaired it The Statistics Department added a new required course Stat 8102 Statistical Methods III to the MS/PhD in Statistics. I designed the curriculum, and offered the course for the first time during Fall We added a new undergraduate minor course Statistics 2521 Data Analysis and Statistical Computing. I designed and offered the course for the first time in Fall