This folder creates a tool to analyze Towed undulating data, especially for Triaxus TUV data.
It can
- Perform automated kriging analysis
- Perform hotspot analysis and cluster analysis
- Visualize the results in 2D and 3D
To use:
source("./src/config.R") # configuration list
source("./src/Triaxus_class.R") # Triaxus base class
source("./src/preprocessing.R")
source("./src/interpolation.R")
source("./src/hotspot.R")
source("./src/clustering.R")
source("./src/plot.R")
source("./src/misc.R")
source("./src/rawFileParser.R")
library(RColorBrewer)
library(leaflet)
library(ggplot2)
options(rgl.useNULL=FALSE)
library(rgl)
myTriaxus <- new("Base_Triaxus",config,newfile,seabirdIndex,bbeIndex,seabird_cutoff,separate=config$separate,rawData = NULL,realName=NULL)
where
- config: list defined in config.R
- newFile: newfile name, can take .dat LOPC rawfile or csv file (sample file can be downloaded here link)
- seabirdIndex, bbeIndex: the index of seabird data and bbe data that represents the same valley points. Need to determine manually. In the newer dataset, the synchronization issues between Seabird and BBE points are solved, so setting seabirdIndex = bbeIndex = 1 should be fine.
- seabird_cutoff: a vector contains the start and end index, for controlling the range of the data to be analyzed.
- separate: Whether to separate upcast and downcast before analyzing
- rawData: Input data to the Triaxus object rather than reading the data. If set, no reading file is performed
- realName: for web app only since the file uploaded may not preserve the original name
To interpolate:
myTriaxus <- preprocessing(myTriaxus) %>% interpolation_main(int_method="krige",det_method="tps") %>% hotspot_main()
"%>%" is the pipeline operation. The preprocessing, interpolation_main and hotspot_main will take Triaxus class as the first input. Different paths can be combined into a list to perform cluster analysis
allTriaxus <- list(myTriaxus) # allTriaxus is a list
# To clustering
# variableForClustering: which variable considered in cluster analysis
# Ks: how many clusters to find
allTriaxus <- clustering_main(allTriaxus, variableForClustering= c("Seabird_temperature","Spec.Cond"),Ks=c(2,3,4,5,6))
# To plot boxplot in each cluster
plot_boxplot (allTriaxus,variables = c("Seabird_Temperature"))
# To plot 3D values, 3D plot take allTriaxus (a list of Triaxus_base class) as input
plot_3d_value(allTriaxus,var="Seabird_Temperature",)
plot_3d_hotspot (allTriaxus,var="Seabird_Temperature")
plot_3d_clustering (allTriaxus,K=2)
# To plot 2D values, 2D plot take Triaxus_base class and variable name
plot_2d(myTriaxus,var= "Seabird_Temperature")
plot_hotspot (myTriaxus,var= "Seabird_Temperature")
plot_raw (myTriaxus,var= "Seabird_Temperature")
If only to parse the raw Triaxus file, there is a function in ./src/rawFileParser.R, readingRawFile. so
source("./src/rawFileParser.R")
df = readingRawFile("triaxusFile.dat") # df is a data frame containing the parsed triaxus data