R/data wrangling/on hold/johnson_2019.R

In chase-lab/BioTIMEx: Aggregating long-term ecological communities data

## johnson_2019
if (FALSE) {
   library(data.table)

   dataset_id <- 'johnson_2019'

   load(file = 'data/raw data/johnson_2019/ddata')
   setDT(ddata)

   setnames(ddata, old = c('Year', 'Creek','Habitat', 'Transect', 'Rep'),
            new = c('year', 'site', 'block', 'plot','subplot'))


   # Wide to long conversion
   species_columns <- grep(colnames(ddata), pattern = 'num.', value = T)
   colnames(ddata) <- trimws( gsub(colnames(ddata), pattern = 'num\\.', replacement = ' ') )
   ddata <- melt(ddata, measure.vars = species_columns, variable.name = 'species', value.name = 'value')
   ddata <- ddata[value > 0] # Deleting useless rows

   # Treatment
   # Typo correction
   ddata[Nutrient == 'Referenceient_Nutrients', Nutrient := 'Reference'][,Nutrient := droplevels(Nutrient)]

   ddata[, treatment := ifelse(Nutrient == "Reference" & Fish == 'Ambient', 'control',
                               ifelse(Nutrient == "Reference" & Fish == 'Low',
                                      'fish manipulation',
                                      ifelse(Nutrient %in% c('Nutrient',"Fertilized") & Fish == 'Ambient',
                                             'nutrient addition',
                                             ifelse(Nutrient %in% c('Nutrient',"Fertilized") & Fish == 'Low',
                                                    'nutrient+fish',
                                                    NA)
                                      )
                               )
   )
   ]

   # Standardisation
   effort <- ddata[, .(effort = length(unique(Month))), by = .(year, site, block, plot, subplot, treatment)]
   ddata <- ddata[, .(value = sum(value)), by = .(site, block, plot, subplot, year, treatment, species)]
   ddata <- merge(ddata, effort, by = c('year', 'site','block', 'plot', 'subplot', 'treatment'))
   # ddata[, value := value / effort]
   # ddata[!is.na(value) & value > 0, value := value / min(value), by = .(year, site, block, plot, subplot, treatment)]      # NOT CHECKED BECAUSE DATASET ON HOLD



   ddata[, ':='(dataset_id = dataset_id,
                timepoints =  paste0('T',seq_along(unique(year))[match(year, unique(year))]),

                design = paste0(ifelse(year == 2003, 'B','A'),
                                ifelse(treatment == 'control', 'C', 'I')),

                time_since_disturbance = year - ifelse(year > 2003 & treatment != 'control', 2004, NA),

                treatment_type = 'eutrophication and community manipulation',
                realm = 'terrestrial',
                taxon = 'invertebrates',
                metric = 'density',
                unit = 'ind/m2',
                comment = 'Block design with two treatments. Treatments were applied at different times and frequencies. Abundances from several samplings a year were pooled together.'
   )
   ][, effort := NULL
   ]


   ddata[, ap:= ifelse(value > 0, 1, 0)][, .(N = sum(ap), S=length(unique(species))), by = .(dataset_id, site, block, plot, subplot, year)][N>S]
   ddata[site == 'West' & block == 'TSA' & plot ==1 & year == 2003]


   dir.create(paste0('data/wrangled data/', dataset_id), showWarnings = FALSE)
   fwrite(ddata, paste0('data/wrangled data/', dataset_id, "/", dataset_id, '.csv'),
          row.names=FALSE)
}