Merge pull request #97 from COMPASS-DOE/viz-metrics-fixes

Viz metrics fixes

Processing_Scripts/sample_catalog_matrix.R

@@ -92,7 +92,8 @@ sample_catalog <- csvs_2 %>%
                             TRUE ~ status)) %>% 
   left_join(metadata_collected, by = c("campaign", "kit_id", "transect_location", "sample_method")) %>% #we need to take into account sample_method, need to know which analytes used which method
   mutate(status = case_when(collected == TRUE & is.na(status) ~ "data available",
-                            TRUE ~ status)) %>% 
+                            TRUE ~ status),
+         transect_location = factor(transect_location, levels = c("upland", "transition", "wetland", "sediment", "water"))) %>% 
   select(-sample_type, -sample_method, -collected, -notes) %>% 
   distinct() %>% 
   pivot_wider(names_from = analyte, values_from = status) %>% 

Processing_Scripts/soil_cec.R

@@ -190,12 +190,14 @@ metadata_collected %>%
 
 data_clean %>% 
   full_join(meta_filter, by = c("campaign", "kit_id", "transect_location")) %>% 
-  mutate(notes = case_when(kit_id == "K050" & transect_location == "upland" ~ "not enough material for extraction",
+  mutate(transect_location = factor(transect_location, levels = c("upland", "transition", "wetland", "sediment")),
+                           notes = case_when(kit_id == "K050" & transect_location == "upland" ~ "not enough material for extraction",
                            kit_id == "K024" & transect_location == "wetland" ~ "sample compromised",
                            collected == FALSE ~ "sample not collected",
                   TRUE ~ notes),
          notes_flags = case_when(is.na(notes) ~ flag,
-                           TRUE ~ notes)) -> full
+                           TRUE ~ notes)) %>% 
+  arrange(kit_id, transect_location) -> full
 
 nums <- sapply(full, is.numeric)           # identify numeric columns
 full[!is.na(full$notes), which(nums)] <- NA  # set compromised kits to NA

Processing_Scripts/soil_iron.R

@@ -214,7 +214,8 @@ data_clean %>%
   full_join(meta_filter, by = c("campaign", "kit_id", "transect_location")) %>% 
   # 2024-03-13: need to split soil and sediments in this script because not all sediments have been ran yet
   filter(sample_type == "soil") %>% 
-  mutate(notes = case_when(kit_id == "K018" & transect_location == "transition" ~ "not enough material for extraction",
+  mutate(transect_location = factor(transect_location, levels = c("upland", "transition", "wetland")),
+         notes = case_when(kit_id == "K018" & transect_location == "transition" ~ "not enough material for extraction",
                            kit_id == "K044" & transect_location == "transition" ~ "not enough material for extraction",
                            kit_id == "K048" & transect_location == "upland" ~ "not enough material for extraction",
                            kit_id == "K050" & transect_location == "upland" ~ "not enough material for extraction",
@@ -223,7 +224,8 @@ data_clean %>%
                            TRUE ~ notes),
          Fe_ug_g = case_when(!is.na(notes) ~ NA,
                              TRUE ~ Fe_ug_g)) %>% 
-  select(campaign, kit_id, transect_location, Fe_ug_g, notes) -> soil_iron
+  select(campaign, kit_id, transect_location, Fe_ug_g, notes) %>% 
+  arrange(kit_id, transect_location) -> soil_iron
 
 # 7. Export L0B data -----------------------------------------------------------
 write_csv(soil_iron, paste0("~/Documents/ec1_soil_iron_L1_", Sys.Date(), ".csv"))

Processing_Scripts/soilsediment_visual_metrics.R

@@ -64,15 +64,25 @@ data_processed <- data_raw %>%
          root_presence = root_mass) %>% 
   dplyr::select(-do_not_use_had_misclassified, -sand_content) %>% 
   mutate(campaign = "EC1",
-         root_thickness = tolower(root_thickness),
          pebbles_angular_rounded = case_match(pebbles_angular_rounded,
-                                              c("angular, rounded", "rounded, angular") ~ "both",
+                                              c("angular, rounded", "rounded, angular") ~ "angular and rounded",
                                               .default = pebbles_angular_rounded),
          root_presence = case_when(grepl("no roots", root_presence) ~ "none",
                                    grepl("few", root_presence) ~ "few",
                                    grepl("moderate", root_presence) ~ "moderate",
                                    grepl("root dominated", root_presence) ~ "many",
-                                   .default = root_presence)) %>% 
+                                   .default = root_presence),
+         uninhabited_shells = case_match(uninhabited_shells, 
+                                         "intact shells" ~ "intact",
+                                         c("intact, fragments", "intact, broken", "broken, intact") ~ "intact and fragments",
+                                         "broken" ~ "fragments",
+                                         .default = uninhabited_shells),
+         visible_minerals = case_match(visible_minerals,
+                                       "quarts" ~ "quartz",
+                                       .default = visible_minerals)) %>% 
+  mutate(root_thickness = case_when(root_presence == "none" ~ NA,
+                                    root_thickness == "Coarse" ~ "fine",
+                                    .default = root_thickness)) %>% 
   relocate(campaign)
 
 #
@@ -83,7 +93,7 @@ cat("Applying flags to", var, "data...")
 ## manipulate a parameter and its flag together
 data_qc <- function(data) {
   data %>% 
-    select(-notes, -cohesion_post_lyopholization) %>% 
+    select(-notes, -cohesion_post_lyopholization, -white_flakes) %>% 
     mutate(
       # switch wetland and transition names due to a...
       # ...sampling error: wetland soil was sampled and put into a jar labeled "transition" incorrectly
@@ -104,7 +114,8 @@ metadata_collected %>%
 data_clean %>%
   full_join(meta_filter, by = c("campaign", "kit_id", "transect_location")) %>% 
   # add rows for samples not collected, creating a "full" dataset of all possible samples
-  mutate(pebbles_angular_rounded = case_when(notes == "kit compromised" ~ NA,
+  mutate(transect_location = factor(transect_location, levels = c("upland", "transition", "wetland", "sediment")),
+         pebbles_angular_rounded = case_when(notes == "kit compromised" ~ NA,
                                              notes == "sample compromised" ~ NA,
                                              TRUE ~ pebbles_angular_rounded),
          root_presence = case_when(notes == "kit compromised" ~ NA,
@@ -128,12 +139,10 @@ data_clean %>%
          glass_plastic = case_when(notes == "kit compromised" ~ NA,
                                    notes == "sample compromised" ~ NA,
                                    TRUE ~ glass_plastic),
-         white_flakes = case_when(notes == "kit compromised" ~ NA,
-                                  notes == "sample compromised" ~ NA,
-                                  TRUE ~ white_flakes),
          notes = case_when(collected == FALSE ~ "sample not collected",
                            TRUE ~ notes)) %>%
-  select(-c(sample_type, sample_method, collected)) -> vizmetrics_full
+  select(-c(sample_type, sample_method, collected)) %>% 
+  arrange(kit_id, transect_location) -> vizmetrics_full
 
 # 6. Write cleaned data to drive -----------------------------------------------