BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260405T103435EDT-7977SCVoiV@132.216.98.100
DTSTAMP:20260405T143435Z
DESCRIPTION:Invited Speaker: Prof. Paul J. Roebber\, Distinguished Professo
 r\, Atmospheric Science Group\, Department of Mathematical Sciences\, Univ
 ersity of Wisconsin at Milwaukee \n\nWhere: Burnside Hall\, Room 934\n\nAb
 stract:\n\nEven my 87-year old mother has heard of “Big Data” and “Data An
 alytics.” But stripping away the marketing noise\, what is it\, really\, a
 nd why should physical scientists such as us care about it? The ability to
  leverage data to improve understanding has always been important\, but is
  becoming increasingly so as data becomes more readily available and  the 
  need  increases  to  extract  some  measure  of  value  from  its  rising
   volume.  Data  analytics  provides  the methodology. The requirements fo
 r a practitioner in this field are application-oriented math and statistic
 s knowledge allied with substantive domain expertise. Since the software t
 ools needed to perform the necessary analyses are not mature\, and often m
 ust be custom-designed\, programming skills are also important.  \n\nMulti
 ple linear regression (MLR) has seen wide use in economics and affiliated 
 fields\, as it is a useful technique for assessing the relationships betwe
 en variables and thereby developing understanding from data. MLR represent
 s an early\, simple application of data analytics to weather prediction in
  the form of Model Output Statistics (MOS)\, which seeks  to  map  numeric
 al  weather  prediction  model  output  to  observations.  More  sophistic
 ated  techniques\,  like  artificial neural networks (ANN)\, including its
  extension to Deep Learning\, or various machine-learning approaches such 
 as Evolutionary Programming\, are now gaining currency in many fields\, an
 d have excellent potential for use in atmospheric sciences. \n\nA straight
 forward example of an atmospheric science question that can be answered wi
 th data analytics is “Can we forecast daily peak electricity load given av
 ailable atmospheric inputs?”  Rather than build a comprehensive\, numerica
 l model that encompasses both the meteorology and the built-environment en
 ergy usage that results\, using data analytics\, we would start by collect
 ing relevant data and building a data model using MLR or and ANN. Given th
 e curse of dimensionality\, which requires an exponential increase in the 
 length of time-series data as the number of variables considered increases
 \, we would need to know something about energy usage to guide our choice 
 of data to collect. The built data model would confirm that the most predi
 ctive variable by far is temperature\, and in the warm season\, apparent t
 emperature (the combination of temperature and humidity)\, but that other 
 information such  as  time-of-day\,  wind  speed and direction\,  cloud  c
 over\,  and  snow  on  the  ground are  also  relevant  in  some situation
 s\, and likewise\, that changing energy usage patterns over time need to b
 e accounted for in the analysis. \n\nA question of interest to a fan of Am
 erican football might be “What is the contribution of penalty calls to NFL
  home field advantage?” Rather than simply argue about it over a beer\, da
 ta analytics can provide an answer. One would collect play-by-play data (a
 vailable online) to build a model of the contribution of factors like posi
 tion on the field\, time remaining in the game\, down-and-distance\, score
 \, and so on to estimate for any situation the win probability. Using that
  model\, we would find that the answer to our original question is approxi
 mately 18%. Data analytics methods are similar in each example\, but the s
 pecifics in each are guided by an understanding of the domain under study.
 \n\nIn this seminar\, I will provide specific examples in the meteorologic
 al domain using MLR\, multiple logistic regression\, ANN\,  and  Evolution
 ary  Programs.  I  will  present  some  future  directions  I  am  develop
 ing\,  including  Deep  Learning applications\, which are highly suited to
  the ubiquitous pattern recognition problems of weather prediction and are
  likely to gain increasing importance in meteorology.\n\nAdditional note: 
 Prof. Roebber is also available to meet with people who are interested on 
 Fri afternoon from 2:30 onwards.  If you are interested\, please contact J
 ohn Gyakum (john.gyakum [at] mcgill.ca).\n\n \n\n \n
DTSTART:20161027T193000Z
DTEND:20161027T203000Z
LOCATION:Room 934\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue 
 Sherbrooke Ouest
SUMMARY:Talk: The Application of Data Analytics to Atmospheric Science
URL:https://www.mcgill.ca/datascience/channels/event/talk-application-data-
 analytics-atmospheric-science-263730
END:VEVENT
END:VCALENDAR