Skip to content
Snippets Groups Projects
Select Git revision
  • master default protected
1 result
Blame Permalink
Code owners
Assign users and groups as approvers for specific file changes. Learn more.
00_CheckData.Rmd 12.31 KiB 
title: 'sPlot 3.0 - Validity Check'
author: "Francesco Maria Sabatini"
output: 
html_document: default
always_allow_html: yes
![](/data/sPlot/users/Francesco/_sPlot_Management/splot-long-rgb.png "sPlot Logo")

Timestamp: r date()
Drafted: Francesco Maria Sabatini
Revised: Stephan Hennekens
Version: 1.3

This report checks for consistency of the dataset used for constructing sPlot 3.0.
Changes to v1.1 - Added check to species data. Created To Do list.
Changes to v1.2 - based on dataset sPlot_3.0.1, received on 29/06/2019 from SH Changes to v1.3 - based on dataset sPlot_3.0.2, received on 24/07/2019 from SH

Key Problems:
The is still a high proportion of plots without location uncertainty

library(reshape2)
library(tidyverse)
library(readr)
library(dplyr)
library(data.table)
library(knitr)
library(kableExtra)
library(viridis)
library(grid)
library(gridExtra)
library(ggforce)
library(xlsx)

Check Header file

Import with parse

header <- readr::read_delim("../sPlot_data_export/sPlot_3_0_2_header.csv", locale = locale(encoding = 'UTF-8'),
                            delim="\t", col_types=cols(
  PlotObservationID = col_double(),
  PlotID = col_double(),
  `TV2 relevé number` = col_double(),
  Country = col_factor(),
  `Cover abundance scale` = col_factor(),
  `Date of recording` = col_date(format="%d-%m-%Y"),
  `Relevé area (m²)` = col_double(),
  `Altitude (m)` = col_double(),
  `Aspect (°)` = col_double(),
  `Slope (°)` = col_double(),
  `Cover total (%)` = col_double(),
  `Cover tree layer (%)` = col_double(),
  `Cover shrub layer (%)` = col_double(),
  `Cover herb layer (%)` = col_double(),
  `Cover moss layer (%)` = col_double(),
  `Cover lichen layer (%)` = col_double(),
  `Cover algae layer (%)` = col_double(),
  `Cover litter layer (%)` = col_double(),
  `Cover open water (%)` = col_double(),
  `Cover bare rock (%)` = col_double(),
  `Height (highest) trees (m)` = col_double(),
  `Height lowest trees (m)` = col_double(),
  `Height (highest) shrubs (m)` = col_double(),
  `Height lowest shrubs (m)` = col_double(),
  `Aver. height (high) herbs (cm)` = col_double(),
  `Aver. height lowest herbs (cm)` = col_double(),
  `Maximum height herbs (cm)` = col_double(),
  `Maximum height cryptogams (mm)` = col_double(),
  `Mosses identified (y/n)` = col_factor(),
  `Lichens identified (y/n)` = col_factor(),
  COMMUNITY = col_character(),
  SUBSTRATE = col_character(),
  Locality = col_character(),
  ORIG_NUM = col_double(),
  ALLIAN_REV = col_character(),
  REV_AUTHOR = col_character(),
  Forest = col_logical(),
  Grassland = col_logical(),
  Wetland = col_logical(),
  `Sparse vegetation` = col_logical(),
  Shrubland = col_logical(),
  `Plants recorded` = col_factor(),
  `Herbs identified (y/n)` = col_factor(),
  Naturalness = col_factor(),
  EUNIS = col_factor(),
  Longitude = col_double(),
  Latitude = col_double(),
  `Location uncertainty (m)` = col_double(),
  Dataset = col_factor(),
  GUID = col_character()
))

This version of sPlot 3.0.1 is composed of r length(unique(header$Dataset)) data sets, for a total of r nrow(header) plots.

Show remaining problems

knitr::kable(problems(header) %>% 
  mutate(Dataset=header$Dataset[problems(header)$row]) %>%
  dplyr::select(Dataset, row, col) %>% 
  distinct(), 
  caption="Problems when importing header data") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), latex_options = "basic", 
                  full_width = F, position = "center")

There seem to be some encoding problems with these plots

Plots without coordinates (by dataset)

knitr::kable(header %>%
  dplyr::select(PlotObservationID, `TV2 relevé number`, Dataset, Longitude, Latitude) %>%
  filter(is.na(Longitude) | is.na(Latitude)) %>% 
  group_by(Dataset) %>%
  summarize(n()), 
  caption="Number of plots without coordinates per dataset") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

Plots without location uncertainty (by dataset).
We could probably assign at least a broad location, with wide uncertainty to the 2020 plots in the Egypt Nile delta dataset.

knitr::kable(header %>% 
               filter(is.na(`Location uncertainty (m)`)) %>% 
               group_by(Dataset) %>% 
               summarize(n()), 
  caption="Number of plots without Location Uncertainty per dataset") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

Big datasets without coordinate uncertainty are:
Czechia_nvd
Germany_vegetweb2
Poland
Slovenia

--> Talk to contributors and ask for 'average' uncertainty (e.g. 100 m)

nNAs <- nrow(header %>% filter(is.na(`Location uncertainty (m)`)))

There are still r nNAs plots without location uncertainty.

Previously known problems:

  1. Import field 'Plants Recorded' into header (SH)
knitr::kable(table(header$`Plants recorded`, exclude=NULL), 
  caption="Number of records for each level in Plants recorded") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

--> Values not available for EVA's datasets. Ask Milan if we can simply assume 'All vascular plants'

  1. Import field 'Herbs identified (y/n)' into header (SH)
knitr::kable(table(header$`Herbs identified (y/n)`, exclude=NULL), 
  caption="Number of records for each level in Plants recorded") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

--> Values not available for EVA's datasets. According to SH, we can simply assume Y.

Check DT table

DT0 <- readr::read_delim("../sPlot_data_export/sPlot_3_0_2_species.csv", 
                            delim="\t", 
                         col_type = cols(
                                PlotObservationID = col_double(),
                                Taxonomy = col_character(),
                                `Taxon group` = col_character(),
                                `Taxon group ID` = col_double(),
                                `Turboveg2 concept` = col_character(),
                                `Matched concept` = col_character(),
                                Match = col_double(),
                                Layer = col_double(),
                                `Cover %` = col_double(),
                                `Cover code` = col_character(),
                                x_ = col_double()
                              )
                         )

Show problems in DT import

knitr::kable(problems(DT0) %>% 
  mutate(Dataset=DT0$Taxonomy[problems(DT0)$row]) %>%
  dplyr::select(Dataset, col, expected, actual) %>% 
  distinct(), 
  caption="Problems when importing Species data") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

id <- as.character(DT0$PlotObservationID[(problems(DT0) %>% dplyr::select(row) %>% distinct())$row])
relnum <- (header %>% filter(PlotObservationID == DT0$PlotObservationID[(problems(DT0) %>% dplyr::select(row) %>% distinct())$row]))$`TV2 relevé number`
db <- (problems(DT0) %>% mutate(Dataset=DT0$Taxonomy[problems(DT0)$row]) %>% dplyr::select(Dataset) %>% distinct())[1,1, drop=T]

All problems seem to be concentrated in PlotID = r id which corresponds to TV2 relevé Number = r relnum in r db.

Other known problems:

#There are some plots without the appropriate cover code
knitr::kable(DT0 %>% 
  filter(`Cover %` ==0  & is.na(`Cover code`)) %>% 
  group_by(Taxonomy) %>% 
  summarize(n()), 
  caption="Summary of DBs without appropriate cover codes") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

# in db British_Columbia_meadows this seems to depend on lichen\moss species
knitr::kable(head(DT0 %>% 
  filter(`Cover %` ==0  & is.na(`Cover code`) & Taxonomy=="British_Columbia_meadows") %>% 
  dplyr::select(PlotObservationID, Taxonomy, `Matched concept`:x_),10), 
  caption="Example from British Columbia meadows db (only first 10 rows)") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

# in db USA_VegBank there seem to be species with only p\a values, next to species with also cover (FMS will have to solve this)
knitr::kable(head(DT0 %>% 
  filter(`Cover %` ==0  & is.na(`Cover code`) & Taxonomy=="USA_VegBank")%>% 
  dplyr::select(PlotObservationID, Taxonomy, `Matched concept`:x_), 10),
  caption="Example from USA_VegBank db (only first 10 rows)") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), 
                  full_width = F, position = "center")

FMS will solve these problems, by assigning an arbitrary low (e.g. 0.1%) cover value to species sampled only as pa. Ideally, FMS should also add a note in the _x field

Distribution of plots across datasets:

knitr::kable(table(header$Dataset), caption="Plots per dataset") %>%
    kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"), full_width = F, position = "center")

Check geographic coordinates

countries <- map_data("world")
ggworld <- ggplot(countries, aes(x=long, y=lat, group = group)) +
  geom_polygon(col=gray(0.3), lwd=0.3, fill = gray(0.9)) +
  theme_bw()  + 
  theme(axis.title = element_blank())

robust.range <- function(x){
  return(c(floor((min(x, na.rm=T)-0.01)/5)*5,
           ceiling((max(x, na.rm=T)+0.01)/5)*5))
}
ggdataset <- list()
tick <- 1
for(d in levels(header$Dataset)){
  datasel <- header %>%
               filter(Dataset==d)
  
  ggdataset[[tick]] <- ggworld + 
    geom_point(data=datasel, aes(x=Longitude, y=Latitude, group=1), 
               col="red", alpha=0.5, cex=1, shape="+") + 
    coord_equal(ylim=robust.range(datasel$Latitude),
                  xlim=robust.range(datasel$Longitude)) + 
    ggtitle(d)
  tick <- tick + 1
  }
  
for(dd in 1:nlevels(header$Dataset)){
  d <- levels(header$Dataset)[dd]
  print(ggdataset[[dd]])
  }
#Depreated below
#Some plots in the Hungary dataset have a altitude >5000 m (!)  

print(ggplot(data=datasel %>% 
           melt()) + 
    geom_histogram(aes(x=value)) + 
    facet_wrap(.~variable, scale="free", ncol=3) + 
    theme_minimal() + 
    theme(axis.text = element_text(size = 8)))
  

#Fix known problems
header.fix <- header %>%
  mutate(`Altitude (m)`=gsub(`Altitude (m)`, pattern=" ", replacement="")) %>%
  mutate(`Altitude (m)`=gsub(`Altitude (m)`, pattern="-", replacement=NA)) %>%
  mutate(`Altitude (m)`=gsub(`Altitude (m)`, pattern="^\\.$", replacement=NA)) %>% 
  mutate(`Altitude (m)`=gsub(`Altitude (m)`, pattern="^\\.", replacement="0\\.")) %>%
  mutate(`Slope (°)`=gsub(`Slope (°)`, pattern="-", replacement=NA)) %>%
  mutate(`Aspect (°)`=gsub(`Aspect (°)`, pattern="-", replacement=NA)) %>%
  mutate(`Lichens identified (y/n)`=replace(`Lichens identified (y/n)`, 
                                            list=`Lichens identified (y/n)` %in% c("0","F", "n", "N" ),
                                            values="FALSE")) %>%
  mutate(`Lichens identified (y/n)`=replace(`Lichens identified (y/n)`, 
                                              list=`Lichens identified (y/n)` %in% c("1","y", "Y"), 
                                              values="TRUE")) %>%
  mutate(`Mosses identified (y/n)`=replace(`Mosses identified (y/n)`, 
                                            list=`Mosses identified (y/n)` %in% c("0","F","f", "n", "N" ),
                                            values="FALSE")) %>%
  mutate(`Mosses identified (y/n)`=replace(`Mosses identified (y/n)`, 
                                            list=`Mosses identified (y/n)` %in% c("1", "j", "J", "T", "y", "Y" ),
                                            values="TRUE"))

write_csv(header.fix, path = "../sPlot_data_export/sPlot_data_header_fix1.csv")