1st draft for drought impact catalog

drought/skill-assessment/drought_catalog/.gitignore

@@ -0,0 +1,8 @@
+.RData
+.Rhistory
+*.csv
+*.png
+*.shp
+*.dbf
+*.prj
+*.shx

drought/skill-assessment/drought_catalog/impactcatalogue_drought.r

@@ -0,0 +1,130 @@
+rm(list=ls())
+library(tidyverse)
+library(lubridate)
+library(sf)
+library(leaflet)
+library(readr)
+library(httr)
+library(sp)
+library(lubridate)
+library(ggplot2)
+library(ggthemes)
+library(cowplot)
+
+
+onedrive_folder <- "c:/Users/pphung/Rode Kruis"
+
+
+## READ ADMIN AND LIVELIHOODZONE BOUNDARIES ---- 
+zwe_lhz <- st_read(sprintf('%s/510 - Data preparedness and IBF - [PRJ] FbF - Zimbabwe - Danish Red Cross/3. Data - Hazard exposure, vulnerability/zwe_livelihoodzones/ZW_LHZ_2011/ZW_LHZ_2011_fixed.shp',
+                           onedrive_folder))
+zwe <- st_read(sprintf('%s/510 - Data preparedness and IBF - [PRJ] FbF - Zimbabwe - Danish Red Cross/3. Data - Hazard exposure, vulnerability/Admin/zwe_admbnda_adm2_zimstat_ocha_20180911/zwe_admbnda_adm2_zimstat_ocha_20180911.shp',
+                       onedrive_folder))
+zwe <- zwe %>%
+  dplyr::select(ADM1_PCODE,ADM2_PCODE,ADM0_EN)
+zwe_lhz <- zwe_lhz %>%
+  dplyr::mutate(ADM0_EN=COUNTRY) %>%
+  dplyr::select(LZCODE)
+
+admin_all <- st_intersection(zwe, zwe_lhz) %>%
+  dplyr::select(ADM1_PCODE,ADM2_PCODE,LZCODE)
+st_geometry(admin_all) <- NULL
+# plot(admin_all)
+
+
+## LOAD AND CALCULATE CROP YIELD ANOMALY [PREDICTANT] ----
+months <- c(1:3,9:12)
+
+yield <- read.csv(sprintf('%s/510 - Data preparedness and IBF - [PRJ] FbF - Zimbabwe - Danish Red Cross/3. Data - Hazard exposure, vulnerability/zwe_cropyield/all_yield_maize_major.csv',onedrive_folder))
+
+mean_sd <- yield %>% 
+  group_by(pcode) %>% 
+  summarise(mean = mean(yield,na.rm=TRUE), sd = sd(yield,na.rm=TRUE)) # calculate mean and standard deviation along the year axis
+yield <- yield %>%
+  left_join(mean_sd,by='pcode')
+yield$yield_anomaly <- (yield$yield-yield$mean)/yield$sd       # calculate CYA
+yield = subset(yield, select=-c(mean,sd))
+
+df_all <- admin_all %>%    # create joint table with adm2, lhz and yield anomaly
+  left_join(yield,by=c('LZCODE'='pcode')) 
+
+df_all <- df_all[rep(seq_len(nrow(df_all)), each = 7), ] %>%
+  group_by(year,LZCODE,ADM2_PCODE) %>%
+  mutate(month = months)
+df_all <- df_all[,c(1,2,3,4,7,5,6)]    # rearrange column "month"
+
+shift <- function(x, n){    # function to shift row up in a column
+  c(x[-(seq(n))], rep(NA, n))
+}
+df_all$yield <- shift(df_all$yield, 4) # shift yield and yield anomaly so that they start from Sept a year before
+df_all$yield_anomaly <- shift(df_all$yield_anomaly, 4)
+
+
+
+
+## LOAD INDICATORS [PREDICTORS] ----
+
+
+## ENSO current ----
+
+enso_zwe <- read.csv(sprintf('%s/510 - Data preparedness and IBF - [RD] Impact-based forecasting/General_Data/elnino/ENSO.csv',onedrive_folder)) %>%
+  # mutate(year=year(Year)) %>%
+  dplyr::select(c(Year,OND))     # choose between "OND"/"NDJ" to fit the impact catalogue
+colnames(enso_zwe)[colnames(enso_zwe) == 'OND'] <- 'enso_cur'
+
+df_all <- df_all %>%       # join to a big table
+  left_join(enso_zwe, by=c('year'='Year'))
+df_all$enso_cur[which(df_all$month %in% c(1:3))] <- NA    # remove enso values in JFM
+
+
+
+## ENSO forecast ----
+
+
+
+
+
+## SPI ----
+
+spi_zwe <- read.csv(sprintf('%s/510 - Data preparedness and IBF - [PRJ] FbF - Zimbabwe - Danish Red Cross/3. Data - Hazard exposure, vulnerability/zwe_spi/zwe_spi3.csv',onedrive_folder))%>%
+  mutate(date=ymd(as.Date(date))) %>%
+  mutate(year=year(date))
+
+df_all <- df_all %>%       # join to a big table
+  left_join(spi_zwe, by=c('LZCODE'='livelihoodzone','year','month')) %>%
+  dplyr::select(-c(X,date))
+
+
+
+
+## SPEI ----
+
+spei_zwe <- read.csv(sprintf('%s/510 - Data preparedness and IBF - [PRJ] FbF - Zimbabwe - Danish Red Cross/3. Data - Hazard exposure, vulnerability/zwe_spi/zwe_spei3.csv',onedrive_folder))%>%
+  mutate(date=ymd(as.Date(date))) %>%
+  mutate(year=year(date))
+
+df_all <- df_all %>%       # join to a big table
+  left_join(spei_zwe, by=c('LZCODE'='livelihoodzone','year','month')) %>%
+  dplyr::select(-c(X,date))
+# write.csv(df_all, './output/drought_catalog.csv')
+
+
+
+
+## Soil moisture ----
+
+
+
+
+
+
+## VCI ----
+
+
+
+
+
+## Cumulative precipitation ----
+
+
+

drought/skill-assessment/drought_catalog/readme.md

@@ -0,0 +1,27 @@
+# Script for creating drought impact catalog
+
+## `impactcatalogue_drought.R`
+
+The script is for arranging predictants (yield, yield anomaly - for maize now) and predictors (drought indicators) per admin2 into a single catalog. 
+
+The arrangement is based on [this excel sheet on Teams ][https://rodekruis.sharepoint.com/sites/510-CRAVK-510/_layouts/15/guestaccess.aspx?guestaccesstoken=5DtgHKRzRrHCPDqpWQl6D1XAJqVfimV3fujFdueQv%2F0%3D&docid=2_027fd1d81d676404d914d155ef55fd16d&rev=1&e=pBm3hv]
+
+The output impact catalog is saved as csv .
+
+### Data input:
+
+The script is now for Zimbabwe, data can be found in the FBF Zimbabwe channel, General_data of IBF channel and can be loaded in easily if synced through OneDrive from the IBF channel in Teams.
+
+At the moment, the inputs are `livelyhood zones`, `admin boundaries`, `crop yield` which are as base. Indicators are now included `ENSO`, `SPI3`, `SPEI3`.
+
+For new indicators, the expected format is similar to one of DMP: in csv, with columns:
+- adm2 pcode and/ or livelihoodzone pcode as same as those in above files
+- date (year, month)
+- values for the indicators
+
+
+### Usage:
+
+Sections `READ ADMIN AND LIVELIHOODZONE BOUNDARIES` and `LOAD AND CALCULATE CROP YIELD ANOMALY [PREDICTANT]` are fixed to define the data frame. Indicators are loaded and modified manually as following:
+- Load the indicator from your synced OneDrive `[indicatorname]_[countryname]`
+- Join to the combined table the indicator based on (1) `livelyhood zones` or `admin boundaries`; (2) `year` and `month` (if `month` available).