In agroimpacts/longisland: Bouy Data Change in Long Island

knitr::opts_chunk$set(echo = TRUE)

next steps (think about all of project along the way) and read about the study to conceptualize water quality http://longislandsoundstudy.net/research-monitoring/water-quality-monitoring/

1. Import CVSs and manipulate as needed (need coordinates and bouy information)

#cheetsheet 
# https://github.com/rstudio/cheatsheets/blob/master/data-transformation.pdf 

install.packages("devtools")
devtools::install_github("tidyverse/dplyr")
library("devtools")
install.packages("utils")
library("utils")
install.packages("tidyverse")
library("tidyverse")
install.packages("dplyr")
library("dplyr")
install.packages("readr")
library("readr")
library(dplyr)
library(tidyr)
library(mapview)

dat <- readr::read_csv("~/Desktop/dummy_dataset.csv")

example

loading data

mastersheet <- readr::read_csv("C:/Users/Sophia/Desktop/ADP/RStudio/sa346/Group Project/LongIsland_Project/Data/CoordinatesTimeDiff.csv")
#the project is my working directoy, and the csv file is in there, so not sure why r is saying it doens't exist 

recognise longitude and latitude, notice each point as Site Name

  read_csv %>% distinct(uuid, .keep_all = TRUE) %>% select(uuid, x, y) %>% 
  mutate(count = 1) %>% st_as_sf(coords = c("x", "y"))

looping to recognise when we upload data again

  dir(pattern = "farmer_spatial.csv|[districts|roads]*.*.shp", 
      full.names = TRUE) %>% 
  map(function(x) {
    if(basename(x) == "farmer_spatial.csv") {
      read_csv(x) %>% st_as_sf(., coords = c("x", "y"), crs = 4326)
    } else {
      st_read(x)
    }
  }) %>% set_names(c("d", "f", "r")) -> sdat

recognise from system file and use as sf object

farmers <- system.file("extdata/farmer_spatial.csv", package = "geospaar") %>% 
  read_csv() %>% st_as_sf(coords = c("x", "y"), crs = 4326)

wired further

library(geospaar)
system.file("extdata/", package = "geospaar") %>% 
  dir(pattern = "farmer_spatial.csv|[districts|roads]*.*.shp", 
      full.names = TRUE) %>% 
  map(function(x) {
    if(basename(x) == "farmer_spatial.csv") {
      read_csv(x) %>% st_as_sf(., coords = c("x", "y"), crs = 4326)
    } else {
      st_read(x)
    }
  }) %>% set_names(c("d", "f", "r")) -> sdat

reprojecting data is

st_transform(sdat$d, crs = st_crs(sdat$r)) %>% 
  st_intersection(x = ., y = rd_sel_buff) %>% st_geometry %>% 
  plot(col = bpy.colors(length(.)))

take a look at points, maybe organize

head()
<- as_tibble()
<- data.table()
aggregate(d$b, list(d$a), mean)
d_tb %>% group_by(a) %>% summarize(mean(b))
d_dt[, mean(b), by = a][order(a)]
#Think about Indexing
a <- 1:10
b <- a
names(b) <- letters[1:10]

mapView

mapview(mastersheet)
addHomeButton(mastersheet)
addMouseCoordinates()
install.packages('ggsn')

upload boey data

statistics on certina fields for individual bouy data

regression on points

graph data, can I make a trend analysis?

regression

time <- 1:nlayers(chirpsz)
fun <- function(x) {
  if(any(is.na(x))) {
    o <- NA
  } else {
    o <- lm(x ~ time)$coefficients[2]  
  }
  o
}
chirpsz_reg <- calc(chirpsz, fun = fun)
chirpsz_reg %>% plot

Mapview

```r

methodology notes

1. graphing and mapping data

Look at class slides

1. stats analysis (will not find significant outputs)

intermediate stats class for pattern/cluster anlaysis - maybe play around with this in SPSS

regression - stats

covariance, Correlation- the degree to which two variables “covary”. Correlation accounts for the standard deviations of the measures and varies from (-1, +1) Correlation coefficient R expresses the degree of fit. Coefficient of Determination (r2) expresses the percentage of variance in Y that is explained by x

Regression- mathematical specification of relationship between two variables, where one is the dependent (Y) variable and one is the independent variable (X)

1. seasonal trend analysis and graphing advacned raster class

Time series analysis

Index Time series Image times series Time Series File

Median trend (Theil-Sen)-slope of every pairwise combination then finds median value. breakdown bond of 29 percent: trends expressed in the image must have persisted for more than 29 percent of their length of the series.

Monotonic trend (Mann-Kendall)- measures how consistently is the trend increasing or decreasing. Has range from -1 to 1, (constantly increases and never decreases and vice versa). A value of 0 indicates no consistent trend.

All pairwise combinations of values over time are evaluated and a tally is made of the number that are increasing or are decreasing over time. Mann-Kendall Significance- produces 2 images, measuring significance of Mann-Kendall Trend one is a significance image expressed as Z scores one is an image that expresses the probability that the observed trend could have occurred by chance.

(there are more notes in google from advaned Raster (Test 1 prep work etc))

agroimpacts/longisland documentation built on May 18, 2019, 12:27 p.m.