benmarwick · November 16, 2024 21:52 · Nov 16, 2024 · Nov 16, 2024 · Nov 16, 2024
diff --git a/prepare-radiocarbon-data-for-analyses-signbase.R b/prepare-radiocarbon-data-for-analyses-signbase.R
@@ -0,0 +1,56 @@
+library(tidyverse)
+library(rcarbon)
+
+signbase <- read_csv("signBase_Version1.0.csv")
+
+# clean up and separate the data in the radiocarbon age columns
+# so we can use it for plotting and analyses. This will generate some
+# warnings, that's ok, nothing to worry about
+signbase_years <- 
+  signbase %>% 
+  drop_na(date_bp_max_min) %>% 
+  mutate(date_bp_max_min = str_replace_all(date_bp_max_min, "\\+\\/\\-", "±")) %>% 
+  mutate(date_bp_max_min = str_replace_all(date_bp_max_min, "\\+", "±")) %>% 
+  separate(date_bp_max_min,
+           sep = " - ",
+           c("date_bp_max", "date_bp_min"),
+           remove = FALSE) %>% 
+  mutate(date_bp_max = ifelse(str_detect(date_bp_max, "\\/"),
+                              str_extract(date_bp_max, ".*?(?=\\/)"),
+                              date_bp_max)) %>% 
+  mutate(date_bp_min = ifelse(str_detect(date_bp_min, "\\/"),
+                              str_extract(date_bp_min, ".*?(?=\\/)"),
+                              date_bp_min)) %>% 
+  separate(date_bp_max,
+           sep = "±",
+           c("date_bp_max_age",
+             "date_bp_max_error")) %>% 
+  separate(date_bp_min,
+           sep = "±",
+           c("date_bp_min_age",
+             "date_bp_min_error")) %>% 
+  drop_na(date_bp_max_age,
+          date_bp_max_error) %>% 
+  mutate(date_bp_max_age = parse_number(date_bp_max_age),
+         date_bp_max_error = parse_number(date_bp_max_error))
+
+# calibrate the radiocarbon ages and join the calibration results back
+# to the main data frame of all the object, location, and sign data.
+# Use this signbase_years_cal data frame for all your anlyses now.
+# Use the column "MedianBP" for visualising and analysing ages. That
+# is calendar years Before the Present
+signbase_years_cal <- 
+  calibrate(signbase_years$date_bp_max_age,
+            signbase_years$date_bp_max_error,
+            ids = signbase_years$object_id,
+            verbose = FALSE) %>% 
+  summary() %>% 
+  tibble() %>% 
+  left_join(signbase,
+            join_by("DateID" == "object_id"))
+
+# for example, we can make a histogram to see which time periods are most
+# frequently represented in the data
+ggplot(signbase_years_cal) +
+  aes(MedianBP) +
+  geom_histogram()
diff --git a/gistfile1.txt → simple-map-of-the-sites-in-SignBase.R b/gistfile1.txt → simple-map-of-the-sites-in-SignBase.R
diff --git a/gistfile1.txt b/gistfile1.txt
@@ -0,0 +1,42 @@
+library(tidyverse)
+library(rnaturalearth)
+library(ggrepel)
+
+# import the data from the paper https://www.nature.com/articles/s41597-020-00704-x
+signbase <- read_csv("signBase_Version1.0.csv")
+
+# prepare simple features objects for a base map of the coastlines
+# and country boundaries of Europe
+world <- ne_countries(scale = "medium", returnclass = "sf")
+Europe <- world[which(world$continent == "Europe"),]
+
+# make a map of the signbase sites!
+ggplot() +
+  # plot the base map of European countries and coastlines
+  geom_sf(data = Europe) +
+  # crop the map to zoom in on the parts of Europe where the sites are
+  coord_sf(xlim = c(-10,30), 
+           ylim = c(35,53), 
+           expand = FALSE) +
+  # add the layer of signbase sites, here I'm using the raw data, but you can 
+  # also use a data frame that you have done some calculations on. You just need
+  # the longitude and latitude columns to be present for this to work
+  geom_point(data = signbase,
+          # point colour indicates 'country', feel free to explore and 
+          # change this to other columns in the signbase_sf data frame
+          aes(x = longitude,
+              y = latitude,
+              # point colour indicates 'country', feel free to explore and 
+              # change this to other columns in the signbase data frame
+              colour = country)) +
+  # add labels for the site names
+  geom_text_repel(data = signbase %>% 
+                    # only one label per site, otherwise it's too croweded to read
+                    distinct(site_name,
+                             .keep_all = TRUE),
+                  aes(x = longitude ,
+                      y = latitude,
+                      label = site_name),
+                  max.overlaps = 100,
+                  size = 2) +
+  theme_minimal()