NextKids pilot embeddings

---

---

young-middle-old

read in embeddings and join in item information

young = read.csv("34_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "young")
middle = read.csv("56_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "middle")
old = read.csv("78_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "old")
adult = read.csv("adult_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "adult")

labs =read.table("78_embeddings/labels.txt", header = FALSE) %>%
      rename(item = V1) %>%
      mutate(item = lapply(str_split(item, ".jpg"), 
                                   function(x) {x[1]}))
d = cbind(labs, young) %>%
  rbind(cbind(labs,middle)) %>%
  rbind(cbind(labs,old)) %>%
  rbind(cbind(labs,adult)) %>%
  rename(x = V1,
         y = V2) %>%
  mutate(item = as.factor(unlist(item))) %>%
  mutate(group = fct_relevel(group, "young", "middle", "old")) 

dict = read.csv("image_dict.csv") %>%
  select(images, type, theme) %>%
  rename(item = images)

d  = left_join(d, dict)

Points

ggplot(d, aes(x = x, y = y, color = type, shape = theme)) +
   geom_point(size = 3) +
   facet_grid(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey", fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        legend.position = "bottom")

Labels

ggplot(d, aes(x = x, y = y)) +
   geom_text(aes(label=item), size = 3) +
   facet_grid(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey", fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank())

median age split

read in embeddings

low = read.csv("low_age_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "low")
high = read.csv("high_age_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "high")


d.m.age= cbind(labs, low) %>%
  rbind(cbind(labs,high)) %>%
  rename(x = V1,
         y = V2) %>%
  mutate(item = as.factor(unlist(item))) %>%
  mutate(group = fct_relevel(group, "low", "high")) %>%
  left_join(dict)

Points

ggplot(d.m.age, aes(x = x, y = y, color = type, shape = theme)) +
   geom_point(size = 3) +
   facet_wrap(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey", fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        legend.position = "bottom")

Labels

ggplot(d.m.age, aes(x = x, y = y)) +
   geom_text(aes(label=item), size = 3) +
   facet_wrap(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey", fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank())

median vocab split

read in embeddings

low = read.csv("low_vocab_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "low")
high = read.csv("high_vocab_embeddings/model.csv", header = FALSE) %>%
  mutate(group = "high")


d.m.vocab = cbind(labs, low) %>%
  rbind(cbind(labs,high)) %>%
  rename(x = V1,
         y = V2) %>%
  mutate(item = as.factor(unlist(item))) %>%
  mutate(group = fct_relevel(group, "low", "high")) %>%
  left_join(dict)

Points

ggplot(d.m.vocab, aes(x = x, y = y, color = type, shape = theme)) +
   geom_point(size = 3) +
   facet_wrap(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey", fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        legend.position = "bottom")

Labels

ggplot(d.m.vocab, aes(x = x, y = y)) +
   geom_text(aes(label=item), size = 3) +
   facet_wrap(~group, scales = "free") +
   theme_bw() +
   theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank(),
        strip.text.x = element_text(size = 14),
        strip.background = element_rect(colour="grey",
                                        fill="grey"),
        axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank())