17. … the parser of every of these elements can easily be changed …
18. myParseSensorML <- function(obj) { root <- xmlRoot(obj) return(xmlName(root)) } mysos = SOS( url = "http://www.sos.de/sos", parsers = SosParsingFunctions( "DescribeSensor" = myParseSensorML) ) … example of replacing a parsing function. Here the parsing for the response to a describesensor operation …
Good that I am first, before the shiny easy user interface for normal people This is what I consider a good user interface... but I don't target normal people, not at all, I target... people like us - Researchers, statisticians, geostatisticians like that! People like R
Png screenshot, blown up, some strange artefacts of the compression Almost like a galaxy, something from outer space, weird, foreign, maybe (!) fascinating, but not really explainable – that is what normal people must think when they use R for the first time My target audience are people who work with R. Not „normal people“, but as the Google chief analyst puts it: The people with the sexiest job in the next ten years!
Personal motivation Something that should be done /is worth starting. A project that is of great use to other people and might actually be used by others Don't want to implement/work with standards for the standards sake only, but use what I learned for interdisciplinary purposes Enable „real“ analysis by real statisticans The power of R (plots, analysis) – > R
Governments organisations
R is a free software environment for statistical computing and graphics. Several platforms, Open Source Statistical modesl, tests, time-series analysis, classification, … you name it! Extensible via packages NO BLACK BOX
SOS 1.0.0 A standardized data storage service for observations made by sensors... Do I even need to explain? It comes with en entourage...
O&M Sampling (SA) OGC Common What is used where? Don't need to read those top to bottom, but good to know what they are about! Capsule this from a user!
What is needed to get R and SOS together? Client and Service Client XML decoding XML encoding KVP encoding XML Rcurl (libcurl)
No XML Beans , some code generating feature in package XML, but couldn't get it to run properly with all the features I need Some design choices in representation of XML in R classes
Flexibility of O&M Some work on simple profile, but disgarded (too big a project to solve „on the side“) Especially: hierachical orderings, optional elements and attributes, GML types with a huge amount of optional metadata attributes DECISION TIME: 52N SOS O&M Profile, not well defined but constructable from code inspection
Interactive encoding of requests and automatic decoding/parsing of responses to R objects for subsequent analysis Never be able to handle all options, so what is the „main feature“?
Transfer method and encoding? Support profile of 52N SOS Interactive creation of GET and POST request Get request as defined by OOSTethys Best Practice We are a community of software developers and marine scientists who develop open source tools to integrate ocean observing systems. Our goal is to dramatically reduce the time it takes to install, adopt and update standards-compliant Web services.
Exchangeability in the code for Parsers , Encoders , Converters Using R feature that methods can be passed as parameter values
Filters not complete! All possible with manual construction... Spatial operators: lacking the encoders for GmlFeatures and construction functions for the features
Filters not complete! All possible with manual construction... Spatial operators: lacking the encoders for GmlFeatures and construction functions for the features Ok so far on the encoding side, BUT I CAN NEVER HANDLE ALL POSSIBILITIES IN A SOS...
R has read.table, read.csv, read... convenience/helping function s Based on user survey Access parts of capabilities
Question for all : Do or do not use SWE/SOS terms??? PRO USING THEM : possible confusing later on, or for experienced users, everybody comes from a certain domain anyway, these are flexible... CON USING THEM : learning threshold, a little bit „unusual“, maybe only for non-native speakers?
Examples! Important to say: Sos4R is NOT for exploration! Non-R users might not get everything From the moment the data frame is given, it's only R !
Bake Z plot(r1clean$Time, r1clean$Wasserstand, type = &quot;l&quot;, ylim=c(200,800))
Based on weathersos
(BETA-) TESTERS WANTED! What is tested? Time series, simple plots What is not tested? Spatial stuff, requests, spatial analysis Support will be given!
Roadmap 1 week: beta RELEASE end user documentation including more tests for coercion and use cases 2 weeks: developer documentation (wiki) and todo list compilation