elsherbini · December 18, 2019 14:41
diff --git a/README.md b/README.md
diff --git a/frequency_of_pizza.png b/frequency_of_pizza.png
diff --git a/pizza_similarity_by_distance.png b/pizza_similarity_by_distance.png
diff --git a/pizzeria_coordinates_and_pope_preferences.csv b/pizzeria_coordinates_and_pope_preferences.csv
diff --git a/pizzeria_menu_item_similarities.csv b/pizzeria_menu_item_similarities.csv
diff --git a/pizzeria_menus.csv b/pizzeria_menus.csv
diff --git a/pope_pizza.png b/pope_pizza.png
diff --git a/pope_pizza.Rmd b/pope_pizza.Rmd
 ---
 title: "R Notebook"
 output: html_notebook
 ---

 ```{r}
 library(tidyverse)
 library(widyr)
 library(cowplot)
 library(ggthemes)
 theme_set(theme_cowplot())

 pizzeria_menus <- read_csv("pizzeria_menus.csv")
 pizzeria_map <- read_csv("pizzeria_coordinates_and_pope_preferences.csv")
 pizzeria_item_similarities <- read_csv("pizzeria_menu_item_similarities.csv")

 p1 <- pizzeria_menus %>% gather("pizza", "presence", -pizzeria) %>% group_by(pizza) %>% summarise(proportion=sum(presence)/n()) %>%
  mutate(pizza=fct_reorder(pizza, proportion)) %>%
  ggplot(aes(y=pizza, x=proportion)) +
  geom_segment(aes(x=0, xend=proportion, y=pizza, yend=pizza))+
  geom_point() +
  scale_y_discrete("Pizza Type") +
  scale_x_continuous("Frequency on menu")

 p2 <- pizzeria_item_similarities %>%
  left_join(pizzeria_map %>% select(-pope) %>%
  gather("coordinate", "value", -pizzeria) %>%
  pairwise_dist(pizzeria, feature = coordinate, value=value) %>%
  select(pizzeria_1=item1, pizzeria_2=item2, distance)) %>%
  left_join(select(pizzeria_map, pizzeria_1=pizzeria, pope_1=pope)) %>%
  left_join(select(pizzeria_map, pizzeria_2=pizzeria, pope_2=pope)) %>%
  mutate(`Pope in common`=if_else(pope_1==pope_2 & !is.na(pope_1) & pope_1 != "none", pope_1, "n/a")) %>%
  mutate(`Pope in common`=fct_relevel(`Pope in common`, "n/a", "Benedict XVI", "Francis", "John Paul II")) %>%
  ggplot(aes(x=distance, y=similarity)) +
  geom_point(aes(color=`Pope in common`), alpha=0.8, size=0.5)+
  #stat_smooth(aes(color=item), method="lm") +
  facet_wrap(~item) +
  scale_x_continuous("Pairwise Geographical Distance", breaks=c(100,300,500)) +
  scale_y_continuous("Pairwise Item Similarity") +
  scale_color_manual(values=c("#999999", "#E69F00", "#0072B2", "#009E73")) +
  ggtitle("","Pizza Similarity by Geographical Distance") +
  theme(strip.text = element_text(size=8))


 p2
 save_plot("frequency_of_pizza.png", p1, base_width=3, base_height=2)
 save_plot("pizza_similarity_by_distance.png", p2, base_width=3.5, base_height=2.5)
 ```

 ```{r}
 # simulating similarities with pope effect
 papal_effect <- tribble(~pope, ~item, ~love,
                        "John Paul II", "mortadella", 2,
                        "Benedict XVI", "diavola", 1,
                        "Benedict XVI", "salsiccia", 1,
                        "Francis", "bistecca", 0.6)

 menu_items <- tribble(~item, ~sacred,
                      "marinara", 0.95,
                      "margherita", 0.95,
                      "diavola", 0.7,
                      "speck", 0.75,
                      "salsiccia", 0.75,
                      "flori_di_zucca", 0.6,
                      "fungo", 0.75,
                      "rapini", 0.95,
                      "bistecca", 0.75,
                      "mortadella", 0.6)

 pairwise_menu_items <- pizzeria_map %>% select(-pope) %>%
  gather("coordinate", "value", -pizzeria) %>%
  pairwise_dist(pizzeria, feature = coordinate, value=value) %>%
  select(pizzeria_1=item1, pizzeria_2=item2, distance) %>%
  select(pizzeria_1, pizzeria_2) %>%
  mutate(item = rep(select(menu_items, item), n())) %>%
  unnest()

 set.seed(42)
 pizzeria_item_similarities <- pizzeria_map %>% select(-pope) %>%
  gather("coordinate", "value", -pizzeria) %>%
  pairwise_dist(pizzeria, feature = coordinate, value=value, upper=FALSE) %>%
  select(pizzeria_1=item1, pizzeria_2=item2, distance) %>%
  mutate(dist_effect = distance/max(distance)) %>%
  left_join(pairwise_menu_items) %>%
  left_join(menu_items) %>%
  left_join(select(pizzeria_map, pizzeria_1=pizzeria, pope_1=pope)) %>%
  left_join(select(pizzeria_map, pizzeria_2=pizzeria, pope_2=pope)) %>%
  left_join(select(papal_effect, pope_1=pope, item, love_1=love)) %>%
  left_join(select(papal_effect, pope_2=pope, item, love_2=love)) %>%
  mutate(similarity = plogis(qlogis(sacred) - dist_effect + rnorm(n(), 0, 0.2) + if_else(!is.na(love_1) & !is.na(love_2) & pope_1==pope_2, love_1, 0))) %>%
  mutate(similarity = if_else(pizzeria_1 == pizzeria_2, 1, similarity)) %>%
  left_join(pizzeria_menus %>% gather(item, presence_1, -pizzeria), by=c("pizzeria_1"="pizzeria", "item"="item")) %>%
  left_join(pizzeria_menus %>% gather(item, presence_2, -pizzeria), by=c("pizzeria_2"="pizzeria", "item"="item")) %>%
  mutate(similarity = if_else(presence_1 == 1 & presence_2 == 1, similarity, NA_real_)) %>%
  select(pizzeria_1, pizzeria_2, pope_1, pope_2, item, similarity)

 pizzeria_item_similarities %>% select(pizzeria_1, pizzeria_2, item, similarity) %>% write_csv("pizzeria_menu_item_similarities.csv")

 pizzeria_item_similarities %>% select(pizzeria_1, pizzeria_2, item, similarity) %>%
  filter(item=="marinara") %>%
  select(-item) %>%
  spread(pizzeria_2, similarity)
  
 pizzeria_item_similarities %>% filter(pizzeria_1 %in% c("P01", "P02"), pizzeria_2 %in% c("P01", "P02"), item=="marinara")
 ```
pizzeria	x	y	pope
P01	85	333	John Paul II
P02	155	350	John Paul II
P03	224	373	John Paul II
P04	128	290	John Paul II
P05	185	300	John Paul II
P06	264	324	none
P07	250	261	none
P08	360	266	Benedict XVI
P09	412	307	Benedict XVI
P10	440	262	Benedict XVI
P11	382	220	Benedict XVI
P12	485	220	none
P13	539	278	none
P14	268	130	none
P15	340	142	Francis
P16	406	136	Francis
P17	312	68	Francis
P18	388	87	Francis
P19	472	98	Francis
P20	558	86	none
pizzeria	marinara	margherita	diavola	speck	salsiccia	flori_di_zucca	fungo	rapini	bistecca	mortadella
P01	1	1	1	1	0	0	1	0	0	1
P02	1	1	1	1	0	0	1	0	0	1
P03	1	1	1	0	1	0	0	0	0	1
P04	1	1	1	1	0	0	1	0	0	1
P05	1	1	1	1	0	0	0	0	0	1
P06	1	1	1	0	0	0	0	0	0	0
P07	1	1	1	1	0	0	1	0	0	0
P08	1	1	1	0	1	0	1	0	0	0
P09	1	1	1	0	1	1	1	0	0	0
P10	1	1	1	1	1	0	1	1	0	1
P11	1	1	1	0	1	0	0	1	0	0
P12	1	1	1	0	0	0	0	1	0	1
P13	1	1	1	1	0	0	1	0	0	1
P14	1	1	1	0	0	1	0	0	0	0
P15	1	1	1	0	0	1	1	0	1	1
P16	1	1	1	0	0	0	0	0	1	0
P17	1	1	1	0	0	0	1	0	1	0
P18	1	1	1	0	1	1	1	0	1	1
P19	1	1	1	0	1	1	1	0	1	1
P20	1	1	1	0	0	1	1	0	0	0
	---
	title: "R Notebook"
	output: html_notebook
	---

	```{r}
	library(tidyverse)
	library(widyr)
	library(cowplot)
	library(ggthemes)
	theme_set(theme_cowplot())

	pizzeria_menus <- read_csv("pizzeria_menus.csv")
	pizzeria_map <- read_csv("pizzeria_coordinates_and_pope_preferences.csv")
	pizzeria_item_similarities <- read_csv("pizzeria_menu_item_similarities.csv")

	p1 <- pizzeria_menus %>% gather("pizza", "presence", -pizzeria) %>% group_by(pizza) %>% summarise(proportion=sum(presence)/n()) %>%
	mutate(pizza=fct_reorder(pizza, proportion)) %>%
	ggplot(aes(y=pizza, x=proportion)) +
	geom_segment(aes(x=0, xend=proportion, y=pizza, yend=pizza))+
	geom_point() +
	scale_y_discrete("Pizza Type") +
	scale_x_continuous("Frequency on menu")

	p2 <- pizzeria_item_similarities %>%
	left_join(pizzeria_map %>% select(-pope) %>%
	gather("coordinate", "value", -pizzeria) %>%
	pairwise_dist(pizzeria, feature = coordinate, value=value) %>%
	select(pizzeria_1=item1, pizzeria_2=item2, distance)) %>%
	left_join(select(pizzeria_map, pizzeria_1=pizzeria, pope_1=pope)) %>%
	left_join(select(pizzeria_map, pizzeria_2=pizzeria, pope_2=pope)) %>%
	mutate(`Pope in common`=if_else(pope_1==pope_2 & !is.na(pope_1) & pope_1 != "none", pope_1, "n/a")) %>%
	mutate(`Pope in common`=fct_relevel(`Pope in common`, "n/a", "Benedict XVI", "Francis", "John Paul II")) %>%
	ggplot(aes(x=distance, y=similarity)) +
	geom_point(aes(color=`Pope in common`), alpha=0.8, size=0.5)+
	#stat_smooth(aes(color=item), method="lm") +
	facet_wrap(~item) +
	scale_x_continuous("Pairwise Geographical Distance", breaks=c(100,300,500)) +
	scale_y_continuous("Pairwise Item Similarity") +
	scale_color_manual(values=c("#999999", "#E69F00", "#0072B2", "#009E73")) +
	ggtitle("","Pizza Similarity by Geographical Distance") +
	theme(strip.text = element_text(size=8))


	p2
	save_plot("frequency_of_pizza.png", p1, base_width=3, base_height=2)
	save_plot("pizza_similarity_by_distance.png", p2, base_width=3.5, base_height=2.5)
	```

	```{r}
	# simulating similarities with pope effect
	papal_effect <- tribble(~pope, ~item, ~love,
	"John Paul II", "mortadella", 2,
	"Benedict XVI", "diavola", 1,
	"Benedict XVI", "salsiccia", 1,
	"Francis", "bistecca", 0.6)

	menu_items <- tribble(~item, ~sacred,
	"marinara", 0.95,
	"margherita", 0.95,
	"diavola", 0.7,
	"speck", 0.75,
	"salsiccia", 0.75,
	"flori_di_zucca", 0.6,
	"fungo", 0.75,
	"rapini", 0.95,
	"bistecca", 0.75,
	"mortadella", 0.6)

	pairwise_menu_items <- pizzeria_map %>% select(-pope) %>%
	gather("coordinate", "value", -pizzeria) %>%
	pairwise_dist(pizzeria, feature = coordinate, value=value) %>%
	select(pizzeria_1=item1, pizzeria_2=item2, distance) %>%
	select(pizzeria_1, pizzeria_2) %>%
	mutate(item = rep(select(menu_items, item), n())) %>%
	unnest()

	set.seed(42)
	pizzeria_item_similarities <- pizzeria_map %>% select(-pope) %>%
	gather("coordinate", "value", -pizzeria) %>%
	pairwise_dist(pizzeria, feature = coordinate, value=value, upper=FALSE) %>%
	select(pizzeria_1=item1, pizzeria_2=item2, distance) %>%
	mutate(dist_effect = distance/max(distance)) %>%
	left_join(pairwise_menu_items) %>%
	left_join(menu_items) %>%
	left_join(select(pizzeria_map, pizzeria_1=pizzeria, pope_1=pope)) %>%
	left_join(select(pizzeria_map, pizzeria_2=pizzeria, pope_2=pope)) %>%
	left_join(select(papal_effect, pope_1=pope, item, love_1=love)) %>%
	left_join(select(papal_effect, pope_2=pope, item, love_2=love)) %>%
	mutate(similarity = plogis(qlogis(sacred) - dist_effect + rnorm(n(), 0, 0.2) + if_else(!is.na(love_1) & !is.na(love_2) & pope_1==pope_2, love_1, 0))) %>%
	mutate(similarity = if_else(pizzeria_1 == pizzeria_2, 1, similarity)) %>%
	left_join(pizzeria_menus %>% gather(item, presence_1, -pizzeria), by=c("pizzeria_1"="pizzeria", "item"="item")) %>%
	left_join(pizzeria_menus %>% gather(item, presence_2, -pizzeria), by=c("pizzeria_2"="pizzeria", "item"="item")) %>%
	mutate(similarity = if_else(presence_1 == 1 & presence_2 == 1, similarity, NA_real_)) %>%
	select(pizzeria_1, pizzeria_2, pope_1, pope_2, item, similarity)

	pizzeria_item_similarities %>% select(pizzeria_1, pizzeria_2, item, similarity) %>% write_csv("pizzeria_menu_item_similarities.csv")

	pizzeria_item_similarities %>% select(pizzeria_1, pizzeria_2, item, similarity) %>%
	filter(item=="marinara") %>%
	select(-item) %>%
	spread(pizzeria_2, similarity)

	pizzeria_item_similarities %>% filter(pizzeria_1 %in% c("P01", "P02"), pizzeria_2 %in% c("P01", "P02"), item=="marinara")
	```