project_presentation_OMS.Rmd

---
title: "Turkey Covid-19 and Vaccination Statistics"
author: "OMS"
date: "21 06 2021"
output:
  xaringan::moon_reader:
    lib_dir: libs
    nature:
      highlightStyle: github
      countIncrementalSlides: no
  html_document:
    df_print: paged
---


# MAT381E

# PROJECT REPORT

# Team OMS

- Semih Çetin



---

# PROJECT INTRODUCTION

+ *In this project is to analyze the spread and vaccine distribution of Covid-19 in Turkey. This virus, which started in 2020 and still poses a life risk with its various variants, has changed our lives in every way.*
+ *With the data I have obtained, I am trying to reveal how many people the virus has made sick in Turkey, how many people have died, how many people have been sick and recovered, and how many people have been vaccinated to protect themselves from this virus.*



---
# PROJECT AIM

+ ***The main aim*** *of the project is to increase the amount of vaccination by finding a relationship between vaccination and the rate of spread of the virus and visualizing it at a level that the public can understand.*

+ ***Another aim*** *is to analyze the dates when the virus peeks and to emphasize that the precaution against the virus should be increased during these dates.*



---
# PROJECT DATA & ACCESS TO DATA

I have used the data of [TURCOVID19](https://turcovid19.com/acikveri/). Worked on 2 Datasets. Both of these datasets are time series datasets.
 
+ *The first Dataset includes Date column and columns with numerical information about the cases.* 
  
+ *Second dataset that used. This data was received via [TURCOVID19](https://turcovid19.com/acikveri/). It contains vaccination data such as total dose, daily dose and total 1st dose by province and date.*


[TUIK](https://data.tuik.gov.tr/Kategori/GetKategori?p=Nufus-ve-Demografi-109) data was used to reach the current population of the provinces.

Geographic location data required to draw maps were obtained from [HUMDATA](https://data.humdata.org/dataset/turkey-administrative-boundaries-levels-0-1-2/resource/6cc83f12-885f-475b-98f7-9bad99d682b9)


---
# TIDYING THE PROJECT DATA

*When I first got Data, there were columns I didn't want. In addition, daily patient data was missing for approximately 260 days. I edited this data by writing a function in Python.This code is as follows:* 

```{R,eval=FALSE,ECHO=FALSE}
import pandas as pd
df=pd.read_excel("data/gunluk_veri.xlsx",
                usecols = "C:P")
for i in range(len(df["Günlük Vaka"])):
    if i==0 :
        df["Günlük Vaka"][i]=df["Toplam Vaka"][i]
    else:
        df["Günlük Vaka"][i]=df["Toplam Vaka"][i]-df["Toplam Vaka"][i-1]
df.to_excel("data/gunluk_veri_son.xlsx")"
```

*Seriously Patient and Entube columns were actually the continuation of each other. The data with NA in the first column were in the second column, and the data with NA in the second column were also in the first column.*

*By combining these columns using the "coalesce" command, a single Serious Patient column was obtained. Unnecessary and missing columns have been removed from the data set.*


---
# Libraries used in the project

```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center'}
library(tidyverse)
library(readxl)
library(forcats)
library(scales)
library(rvest)
library(sf)
library(viridis)
library(plotly)
```

+ tidyverse
+ readxl
+ forcats
+ scales
+ rvest
+ sf
+ viridis
+ plotly

---


# ABOUT THE PROJECT DATA

The Covid-19 Disease dataset consists of 9 columns (But 8 of them were used.) and 645 lines. Stored as *data*

```{r message= F, warning=FALSE,results = 'hide', out.height='100%', out.width='50%',  fig.align='center'}
data=read_excel("data/gunluk_veri_son.xlsx",range="B1:O646")
data=as_tibble(data)

data=data%>%
  rename(Agır_hasta="Ağır Hasta")%>%
  mutate(Ağır_Hasta= coalesce(Entube,Agır_hasta))%>%
  select(-c("Toplam_Test","YBU","Entube","Agır_hasta","Toplam_Hasta","Gunluk_Hasta"))
head(data)
```

---
### The contents of the Covid-19 Disease Data :

+ Date (Y-M-D)
+ Total Case
+ Daily Case
+ Total Death
+ Daily Death
+ Total Recovery
+ Daily Recovery
+ Serious Patient


---


The Covid-19 Vaccination dataset consists of 10 columns (But 5 of them were used.) and 19129 lines. Stored as *data_vac

```{r,warning=FALSE,results = 'hide'}

data_vac=read_excel("data/gunluk_asi.xlsx",range="C1:L19130")
data_vac=as_tibble(data_vac)
data_vac$Tarih<-as.POSIXct(format(as.POSIXct(data_vac$Tarih,format='%m/%d/%Y %H:%M:%S'),format='%Y-%m-%d'))
class(data_vac$Tarih)
head(data_vac)
```


### The contents of the Covid-19 Vaccination Data :

+ Province
+ Date (Y-M-D)
+ Total Dose
+ Total First Dose
+ Daily First Dose



---
#Data Visualization

*Line graph was used for Case Data visualization.*

*Desired data was filtered and "Daily Case, Daily Death, Daily Recovered Total Case, Total Death, Total Serious Patient and Total Recovered " line graphs were created. The graphics were made interactive using the Plotly library.*


```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center',results = 'hide'}
options(scipen = 999) # turn off scientific notation such as 4e+25
Sys.setlocale("LC_TIME", "UK") #This code is required so that the month data on the x-axis is not Turkish in the charts.
```

```{r,warning=FALSE}
line_gunluk=data%>%
  select("Tarih", "Günlük Vaka")%>%
  rename(Daily_Case="Günlük Vaka",Date="Tarih") %>%
  ggplot(aes(x=as.Date(Date),y=Daily_Case))+
  geom_line(stat="identity",color=("#8E1803"))+
  labs(x = "Date", y = "Daily Case",
    title = "Number of daily coronavirus cases in Turkey")+
  scale_x_date(breaks=date_breaks("3 months"),
      labels=date_format("%b %y"))+
  theme_minimal()

```


---

```{r echo=F, message= F, warning=FALSE,  fig.align='center'}
ggplotly(line_gunluk)
```

>The daily case count is peeking in December 2020, April 2021 and November 2021. 


---

```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center',results = 'hide'}
line_vefatg=data%>%
  select("Tarih", "Gunluk_Vefat")%>%
  rename(Daily_Death="Gunluk_Vefat",Date="Tarih") %>%
  ggplot(aes(x=as.Date(Date),y=Daily_Death))+
  geom_line(stat="identity")+
  labs(x = "Date", y = "Daily Death",
    title = "Number of daily coronavirus deaths in Turkey")+
  scale_x_date(breaks=date_breaks("2.5 months"),
      labels=date_format("%b %y"))+
  theme_minimal()


```

```{r echo=F, message= F, warning=FALSE,  fig.align='center'}
ggplotly(line_vefatg)
```

>Since the number of cases is directly related to the number of deaths and recovery, other variables are also peeking at these dates. 

---
```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center',results = 'hide'}
line_iyileseng=data%>%
  select("Tarih", Gunluk_iyilesen)%>%
  rename(Daily_recovery="Gunluk_iyilesen",Date="Tarih")%>%
  drop_na()%>%
  ggplot(aes(x=as.Date(Date),y=Daily_recovery))+
  geom_line(stat="identity",color=("#0E6655"))+
  labs(x = "Date", y = "Daily Recoveries",
    title = "Number of daily coronavirus recoveries in Turkey")+
  scale_x_date(breaks=date_breaks("2.5 months"),
      labels=date_format("%b %y"))+
  theme_minimal()
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(line_iyileseng)
```

>In these line graphs , we can observe that the graphs make exactly the same movement. As a result, the correlation is very high.

---
*Line graph was also used to show values such as total cases.*

```{r message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center'}
line_toplam=data%>%
  select("Tarih", "Toplam Vaka")%>%
  rename(Total_Case="Toplam Vaka",Date="Tarih") %>%
  ggplot(aes(x=as.Date(Date),y=Total_Case))+
  geom_line(stat="identity",color=("#8E1803"))+
  labs(x = "Date", y = "Total Case",
    title = "Number of Total coronavirus cases in Turkey")+
  scale_x_date(breaks=date_breaks("3 months"),
      labels=date_format("%b %y"))+
  theme_minimal()
```


---
```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(line_toplam)
```


> As expected, the Total Cases, Deaths, and Recovery numbers are progressing cumulatively.

---

*Line graph, Bar Graph and Map Visualization were used for Covid 19 Vaccination Data Visualization.*

```{r message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center'}
line_asıg= data_vac %>%
  select("İl","Tarih","Günlük 1. Doz")%>%
  rename(First_Dose="Günlük 1. Doz",Date="Tarih")%>%
  filter(İl=="Türkiye")%>%
  ggplot(aes(x=as.Date(Date),y=First_Dose))+
  geom_line(stat="identity",colour="#2980B9")+
  labs(x = "Date", y = "First Dose",
    title = "Number of daily first dose of coronavirus vaccination in Turkey")+
  scale_x_date(breaks=date_breaks("1 months"),
      labels=date_format("%m %y"))+
  theme_minimal()
```


---
```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(line_asıg)
```


>*Vaccination occurs at a normal rate until June 2021, very quickly from July 2021 to October 2021, and very slowly after November 2021.* 
>*It was observed that the number of daily deaths decreased rapidly in June 2021, when daily vaccination peaked.*

---

*Provinces were divided according to their regions. The daily and total vaccination rates of the regions were compared.*

```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center',results = 'hide'}
marmara = list("Istanbul" , "Balıkesir", "Bursa", "Tekirdağ", "Çanakkale",
          "Yalova", "Kocaeli", "Kırklareli", "Edirne", "Bilecik", "Sakarya")

ege = list("Izmir", "Manisa", "Aydın", "Denizli", "Uşak", "Afyon", "Kütahya", "Muğla")

ic_anadolu = list("Ankara", "Konya", "Kayseri", "Eskişehir", "Sivas", "Kırıkkale", "Aksaray",
             "Karaman", "Kırşehir", "Niğde", "Nevşehir", "Yozgat", "Çankırı")

karadeniz = list("Amasya", "Artvin", "Bartın", "Bayburt", "Bolu", "Çorum","Düzce", "Gümüşhane", "Giresun",
            "Karabük", "Kastamonu", "Ordu", "Rize", "Samsun", "Sinop", "Tokat", "Trabzon", "Zonguldak")

dogu_anadolu = list("Ağrı", "Ardahan", "Bitlis", "Bingöl", "Elazığ", "Erzincan", "Erzurum",
               "Hakkari", "Iğdır", "Kars", "Malatya", "Muş","Tunceli", "Van")

gdogu_anadolu = list("Gaziantep", "Diyarbakır", "Şanlıurfa", "Batman", "Adıyaman",
                "Siirt", "Mardin", "Kilis", "Şırnak")

akdeniz = list("Antalya", "Adana", "Mersin", "Hatay", "Burdur", "Osmaniye",
          "Kahramanmaraş", "Isparta","Içel")
```

```{r echo=F, message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center',results = 'hide'}
data_vacbolge=data_vac %>%
  mutate(Bölge=case_when(
    data_vac$İl %in% marmara ~ "Marmara",
    data_vac$İl%in%ege ~ "Ege",
    data_vac$İl%in%ic_anadolu ~ "Ic_Anadolu",
    data_vac$İl%in%karadeniz ~ "Karadeniz",
    data_vac$İl%in%dogu_anadolu ~ "Dogu_Anadolu",
    data_vac$İl%in%gdogu_anadolu ~ "Gdogu_Anadolu",
    data_vac$İl%in%akdeniz ~ "Akdeniz")) %>%
  group_by(Bölge,Tarih)%>%
  drop_na(Bölge)%>%
  rename(Birinci_Doz="Toplam 1. Doz",Region="Bölge",Date="Tarih")%>%
  summarise(First_dose=sum(Birinci_Doz),.groups="keep")%>%
  ggplot(aes(x=as.Date(Date),y=First_dose, color=Region))+
  geom_line(stat="identity")+
  labs(x = "Date", y = "Total First Dose",
       title = "Number of total first dose of coronavirus vaccination in 7 Region")+
  scale_x_date(breaks=date_breaks("1 months"),
      labels=date_format("%m %y"))+
  theme_minimal()
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(data_vacbolge)
```

---

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
data_vacbolge=data_vac %>%
  mutate(Bölge=case_when(
    data_vac$İl %in% marmara ~ "Marmara",
    data_vac$İl%in%ege ~ "Ege",
    data_vac$İl%in%ic_anadolu ~ "Ic_Anadolu",
    data_vac$İl%in%karadeniz ~ "Karadeniz",
    data_vac$İl%in%dogu_anadolu ~ "Dogu_Anadolu",
    data_vac$İl%in%gdogu_anadolu ~ "Gdogu_Anadolu",
    data_vac$İl%in%akdeniz ~ "Akdeniz")) %>%
  group_by(Bölge,Tarih)%>%
  drop_na(Bölge)%>%
  rename(Birinci_Doz="Günlük 1. Doz",Region="Bölge",Date="Tarih")%>%
  summarise(First_dose=sum(Birinci_Doz),.groups="keep")%>%
  ggplot(aes(x=as.Date(Date),y=First_dose, color=Region))+
  geom_line(stat="identity")+
   facet_grid(Region ~ .)+
  labs(x = "Date", y = "Daily First Dose",
      title = "Number of daily first dose of coronavirus vaccination in 7 Region")+
  scale_x_date(breaks=date_breaks("1 months"),
      labels=date_format("%m %y"))+
  theme_minimal()
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(data_vacbolge)
```

> As shown in the figure, Marmara is the region with the most vaccination, and Eastern Anatolia is the region with the least vaccination.

---
*The province with the most and at least 5 vaccinations on a provincial basis was shown with a bar graph.*

```{r message= F, warning=FALSE, out.height='100%', out.width='50%',  fig.align='center'}
data_vacil_h=data_vac %>%
  select("İl","Tarih","Toplam 1. Doz") %>%
  filter(İl!=c("Türkiye")& Tarih=="2021-09-13")%>%
  rename(First_dose="Toplam 1. Doz",Province="İl")%>%
  arrange(First_dose)%>%
  tail(5)%>%
  ggplot(aes(x=Province,y=First_dose))+
  geom_bar(stat="identity",fill="#1E8449")+
  labs(x = "City", y = "Total First Dose",
    title = "5 provinces with the highest number of vaccinations")+
  theme_minimal()
```



---

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(data_vacil_h)
```


> As shown in figure, Istanbul is the province with the highest number of vaccinations in Turkey.
---

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
data_vacil_l=data_vac %>%
  select("İl","Tarih","Toplam 1. Doz") %>%
  filter(İl!=c("Türkiye") & Tarih=="2021-09-13")%>%
  rename(First_dose="Toplam 1. Doz",Province="İl")%>%
  arrange(desc(First_dose))%>%
  tail(5)%>%
  ggplot(aes(x=Province,y=First_dose))+
  geom_bar(stat="identity",fill="#A93226")+
  labs(x = "City", y = "Total First Dose",
    title = "5 provinces with the least vaccination")+
  theme_minimal()
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
ggplotly(data_vacil_l)
```


> As shown in figure, Bayburt is the province with the lowest number of vaccinations in Turkey.


---

*Turkey's provincial border map was taken over [HUMDATA](https://data.humdata.org/dataset/turkey-administrative-boundaries-levels-0-1-2/resource/6cc83f12-885f-475b-98f7-9bad99d682b9) to be used as spatial data. Stored as **turkey**.* 

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
map_data= data_vac %>%
  select("İl","Toplam 1. Doz") %>%
  filter(İl!=c("Türkiye"))%>%
  rename(Birinci_Doz="Toplam 1. Doz",adm1_tr="İl")%>%
  group_by(adm1_tr)%>%
  summarise(İl_bdoz=sum(Birinci_Doz),.groups="keep")
```

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
map_data$adm1_tr=str_to_upper(map_data$adm1_tr,locale="tr")
```

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
map_data$adm1_tr[map_data$adm1_tr=="AFYON"]="AFYONKARAHİSAR"
map_data$adm1_tr[map_data$adm1_tr=="ESKİŞEHİR"]="ESKİŞEHIR"
map_data$adm1_tr[map_data$adm1_tr=="IÇEL"]="MERSİN"
map_data$adm1_tr[map_data$adm1_tr=="ISTANBUL"]="İSTANBUL"
map_data$adm1_tr[map_data$adm1_tr=="IZMİR"]="İZMİR"

nufus=read_excel("data/il_nufus.xls",range="A6:B86",col_names = F)
colnames(nufus)=c("adm1_tr","nufus")
nufus=as_tibble(nufus)
nufus$adm1_tr=str_to_upper(nufus$adm1_tr,locale="tr")
nufus$adm1_tr[nufus$adm1_tr=="ESKİŞEHİR"]="ESKİŞEHIR"
map_data=full_join(map_data,nufus,by="adm1_tr")

map_data$oran=map_data$İl_bdoz/map_data$nufus
map_data$oran=round(map_data$oran,digits=1)
```

```{r,warning=FALSE,message=FALSE,results = 'hide'}
turkey <- st_read("data/tur_polbnda_adm1.shp")
```

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
turkey$adm1_tr[map_data$adm1_tr!=turkey$adm1_tr]
map_data$adm1_tr[map_data$adm1_tr!=turkey$adm1_tr]
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
plot(st_geometry(turkey))
```


---
*The names of the provinces were arranged as Turkish characters and capital letters both in the Spatial Data and in the Vaccination Data. Two datasets were joined by using the **full_join** *method by using "adm1_tr" column.*
```{r}
map_last=full_join(map_data,turkey,by="adm1_tr")
```

---

*The combined data was used to build the Static Map by using ggplot.Label showing the name of the province and the vaccination rate was attached. *

```{r,warning=FALSE,message=FALSE}
colors=c("#d7263d","#f46036","#F4D03F","#95e06c")
breaks=c(0,55,65,70)
labels <-  sprintf("%s\n%s", map_last$adm1_tr, map_last$oran)

map_last1=map_last%>%
  ggplot()+
  geom_sf(aes(fill=oran,geometry=geometry),color="black")+
  labs(title = "Total 1st Dose Vaccination by Province",
       caption = "Data Source: turcovid19.com")+
  theme_void() +
  theme(title = element_text(face="italic"))+
  geom_sf_label(aes(label = labels, geometry = geometry, size=1), colour = "#6E2C00", size=1.8,fill="snow")+
  scale_fill_stepsn(colors=colors,breaks=breaks)

```

---

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
map_last1

```

> As shown in the figure, There is a visible link between level of education , number of books read and rate of vaccination. Naturally, the rate of vaccination increases as you move from east to west.


---
*Dynamic Map created using GGplot was made interactive using the Plotly library. But i couldn't render it on Presentation. So want to show you map without info label.*

```{r echo=F, message= F, warning=FALSE, out.height='50%', out.width='100%',  fig.align='center',results = 'hide'}
colors=c("#d7263d","#f46036","#F4D03F","#95e06c")
breaks=c(0,55,65,70)

map_last2=map_last%>%
  ggplot()+
  geom_sf(aes(fill=oran,geometry=geometry),color="black")+
  labs(title = "Total 1st Dose Vaccination by Province",
       caption = "Data Source: turcovid19.com")+
  theme_void() +
  theme(title = element_text(face="italic"))+
  geom_sf_label(aes(label = oran, geometry = geometry, size=1), colour = "#6E2C00", size=1.8,fill="snow")+
  scale_fill_stepsn(colors=colors,breaks=breaks)
```

```{r echo=F, message= F, warning=FALSE,fig.align='center'}
map_last2

```

---

# Conclusion

*As a result, there is a high correlation between Turkey Covid-19 and Vaccination. As vaccination increases, the number of cases and the lethality of the disease decrease. Vaccination rates are higher in provinces with higher education levels. We must be vaccinated to reduce the lethal effect of the virus. The measures taken against the virus should be increased when the number of cases peaks. On November 13, 2021, the last date for which data is available, the highest number of vaccinations is in Istanbul, and the lowest in Bayburt. On a regional basis, Marmara has the highest number of vaccinations, while Eastern Anatolia has the lowest.*

---
#References

[https://turcovid19.com/acikveri/](https://turcovid19.com/acikveri/)

[https://ourworldindata.org/coronavirus](https://ourworldindata.org/coronavirus)

[https://data.humdata.org/dataset/turkey-administrative-boundaries-levels-0-1-2/resource/6cc83f12-885f-475b-98f7-9bad99d682b9](https://data.humdata.org/dataset/turkey-administrative-boundaries-levels-0-1-2/resource/6cc83f12-885f-475b-98f7-9bad99d682b9)

[https://data.tuik.gov.tr/Kategori/GetKategori?p=Nufus-ve-Demografi-109](https://data.tuik.gov.tr/Kategori/GetKategori?p=Nufus-ve-Demografi-109)

[https://chartio.com/learn/charts/how-to-choose-colors-data-visualization/](https://chartio.com/learn/charts/how-to-choose-colors-data-visualization/)