sxh_v2.Rmd

---
title: "Kết quả SXH"
output:
  word_document: default
  html_document: default
echo: false
warning: false
message: false
---

```{r setup, include=FALSE}
library(flextable)
library(officer)
library(ggplot2)
library(ggrepel)
library(tidyverse)

knitr::opts_chunk$set(echo = FALSE)
```

# Đợt dịch năm 2019

## Thành phố 

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_tp_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r, dpi=500}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Outbreak 2019") +
  scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("Unqualified", "High sens",
                                "High spec", "Both high")) +
  scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 1

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q1_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q1 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D1") +
  scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
  scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 3

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q3_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q3 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D3") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 4

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q4_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q4 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D4") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 5

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q5_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q5 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D5") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 6

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q6_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q6 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D6") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 7

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q7_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q7 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D7") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 8

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q8_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q8 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D8") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 10

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q10_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q10 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D10") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 11

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q11_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q11 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D11") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 12

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_q12_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q12 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D12") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Gò Vấp

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_gv_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
gv <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Go Vap D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Tân Bình

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_tb_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tb <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Tan Binh D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Thạnh

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_bt_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bt <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Thanh D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Phú Nhuận

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_pn_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
pn <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Phu Nhuan D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Củ Chi

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_cc_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
cc <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Cu Chi SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Hóc Môn

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_hm_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
hm <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Hoc Mon D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Chánh

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_bc_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bc <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Chanh SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Nhà Bè

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_nb_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
nb <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Nha Be SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Thủ Đức

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_tptd_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tptd <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Thu Duc City") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Cần Giờ

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_cg_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
cg <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Can Gio SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Tân Phú

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_tphu_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tphu <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Tan Phu D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Tân

```{r}
tbl <- readRDS("Data/obs_sxh_2019/tbl_bta_sxh_2019.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bta <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Tan D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

```{r, fig.width=14, fig.height=14, dpi=500}
library(patchwork)

wrap_plots(list(q1, q3, q4 , q5), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(q6, q7, q8 , q10), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(q11, q12, gv , tb), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(bt, pn, cc , hm), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(bc, nb, tptd, cg), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")
```

```{r, fig.width=14, fig.height=7, dpi=500}
wrap_plots(list(tphu, bta), ncol = 2, nrow = 1) + 
  theme(legend.position = "bottom")
```

# Đợt dịch năm 2022

## Thành phố 

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_tp_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r, dpi=500}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Outbreak 2022") +
  scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("Unqualified", "High sens",
                                "High spec", "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 1

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q1_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q1 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D1") +
  scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 3

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q3_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q3 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D3") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 4

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tbl <- readRDS("Data/obs_sxh_2022/tbl_q4_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q4 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D4") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 5

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q5_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q5 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D5") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 6

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q6_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q6 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D6") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 7

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q7_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q7 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D7") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 8

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q8_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q8 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D8") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 10

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q10_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q10 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D10") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 11

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q11_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q11 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D11") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Quận 12

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_q12_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
q12 <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "D12") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Gò Vấp

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_gv_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
gv <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Go Vap D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Tân Bình

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_tb_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tb <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Tan Binh D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Thạnh

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_bt_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bt <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Thanh D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Phú Nhuận

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_pn_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
pn <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Phu Nhuan D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Củ Chi

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_cc_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
cc <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Cu Chi SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Hóc Môn

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_hm_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
hm <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Hoc Mon D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Chánh

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_bc_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bc <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Chanh SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Nhà Bè

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_nb_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
nb <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Nha Be SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Thủ Đức

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_tptd_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tptd <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Thu Duc City") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Cần Giờ

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_cg_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
cg <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Can Gio SD") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Tân Phú

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_tphu_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
tphu <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Tan Phu D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

## Bình Tân

```{r}
tbl <- readRDS("Data/obs_sxh_2022/tbl_bta_sxh_2022.rds")

tbl <- tbl %>%
  mutate(cat = case_when(
    sens >= 0.8 & spec >= 0.5 & spec < 0.8 ~ 1,
    spec >= 0.8 & sens >= 0.5 & sens < 0.8 ~ 2,
    sens >= 0.8 & spec >= 0.8 ~ 3,
    TRUE ~ 0
  ))

tbl$cat <- as.factor(tbl$cat)
tbl <- tbl %>%
  mutate(across(where(is.numeric), ~ sprintf("%.3f", .)))
my_table <- flextable(tbl)

my_table <- theme_vanilla(my_table)
my_table <- autofit(my_table)
my_table
```

```{r}
tbl$sens <- as.numeric(tbl$sens)
tbl$fpr <- as.numeric(tbl$fpr)
bta <- ggplot(tbl, aes(x = fpr, y = sens)) +
  geom_point(aes(color = cat), size = 3) +
  geom_text_repel(aes(label = method), vjust = -0.5, max.overlaps = Inf) +
  labs(x = "False Positive Rate",
       y = "Sensitivity",
       color = NULL,
       title = "Binh Tan D") +
    scale_color_manual(values = c("0" = "lightgrey", "1" = "blue",
                                "2" = "green", "3" = "red"),
                     labels = c("0" = "Unqualified", 
                                "1" ="High sens",
                                "2" = "High spec", 
                                "3" = "Both high")) +
    scale_x_continuous(breaks = seq(0,1,0.2)) +
  scale_y_continuous(breaks = seq(0,1,0.2)) +
  theme_classic() +
  theme(legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
```

```{r, fig.width=14, fig.height=14, dpi=500}
library(patchwork)

wrap_plots(list(q1, q3, q4 , q5), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(q6, q7, q8 , q10), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(q11, q12, gv , tb), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(bt, pn, cc , hm), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")

wrap_plots(list(bc, nb, tptd , cg), ncol = 2, nrow = 2) + 
  theme(legend.position = "bottom")
```

```{r, fig.width=14, fig.height=7, dpi=500}
wrap_plots(list(tphu, bta), ncol = 2, nrow = 1) + 
  theme(legend.position = "bottom")
```