DV_project.Rmd

---
title: "Group U_Final project"
author: "shiyun li, xia shan, junjie ma, peiyuan huo"
date: "4/25/2022"
output:
  html_document: default
  pdf_document: default
---

# Data Preparing:

- The dataset is from LA opendate, containing crimes happened during 2020-2022, updated weekly by LAPD
- Delete the cases that are still under investigation, and cleaned the text
- Integragate crime category that happened more than 500 times in the past

```{r, echo=FALSE, warning=FALSE, message = FALSE}
library(readr)
library(DT)
library(dplyr)
crime=read.csv("/Users/kimberlyshan/Desktop/Crime_Data_from_2020_to_Present.csv")
```


```{r, echo=FALSE,warning=FALSE}
#Clean version of the total dataset
col=c("DR_NO","Date.Rptd","Date.occ","TIME.OCC","AREA","AREA.NAME","Rpt.Dist.No","Crm.Cd.Desc","Mocodes",
      "Vict.Age","Vict.Sex","Vict.Descent","Premis.Desc","Weapon.Desc","Status.Desc","LOCATION","LAT","LON")
clean_total=crime[,colnames(crime) %in% col]

clean_total$TIME.OCC <- as.character(clean_total$TIME.OCC) 
proper_case <- function(x) {
  return (gsub("\\b([A-Z])([A-Z]+)", "\\U\\1\\L\\2" , x, perl=TRUE))
}
clean_total <- clean_total %>% mutate(Crm.Cd.Desc = proper_case(Crm.Cd.Desc),
                    AREA=proper_case(AREA),
                    AREA.NAME=proper_case(AREA.NAME),
                    LOCATION=proper_case(LOCATION),
                    Weapon.Desc = proper_case(Weapon.Desc),
                    Premis.Desc=proper_case(Premis.Desc),
                    Status.Desc=proper_case(Status.Desc))
```


```{r, echo=FALSE,warning=FALSE}
#Delete the cases that are still under investigation
sum_ctg=clean_total %>% 
  group_by(Status.Desc)%>%
  summarise(length(Status.Desc))

total1=subset(clean_total,Status.Desc=="Adult Arrest"|Status.Desc=="Adult Other" |Status.Desc=="Juv Arrest"|Status.Desc=="Juv Other")
```


```{r, echo=FALSE,warning=FALSE}
#Integragate crime category
sum_crime=total1 %>% 
  group_by(Crm.Cd.Desc)%>%
  summarise(length(Crm.Cd.Desc))
#select the cirmes that happen more than 500 times 
clean_total=subset(total1,Crm.Cd.Desc=="Intimate Partner - Simple Assault"|Crm.Cd.Desc=="Battery - Simple Assault"|Crm.Cd.Desc=="Assault With Deadly Weapon, Aggravated Assault
"|Crm.Cd.Desc=="	Vandalism - Felony ($400 & Over, All Church Vandalisms) "|Crm.Cd.Desc=="Intimate Partner - Aggravated Assault"|Crm.Cd.Desc=="Criminal Threats - No Weapon Displayed"|Crm.Cd.Desc=="Robbery"|Crm.Cd.Desc=="Vehicle-Stolen"|Crm.Cd.Desc=="Burglary"|Crm.Cd.Desc=="Violation Of Restraining Order "|Crm.Cd.Desc=="Brandish Weapon "|Crm.Cd.Desc==" Vandalism - Misdeameanor ($399 Or Under)"|Crm.Cd.Desc=="Violation Of Court Order "|Crm.Cd.Desc==" Theft Plain - Petty ($950 & Under)"|Crm.Cd.Desc==" Letters, Lewd  -  Telephone Calls, Lewd"|Crm.Cd.Desc=="Child Abuse (Physical) - Simple Assault "|Crm.Cd.Desc=="Theft-Grand ($950.01 & Over)Excpt,Guns,Fowl,Livestk,Prod "|Crm.Cd.Desc==" Trespassing"|Crm.Cd.Desc=="Other Miscellaneous Crime "|Crm.Cd.Desc=="Contempt Of Court "|Crm.Cd.Desc==" Shoplifting - Petty Theft ($950 & Under)"|Crm.Cd.Desc==" Attempted Robbery"|Crm.Cd.Desc==" Battery Police (Simple)
"|Crm.Cd.Desc=="Battery With Sexual Contact "|Crm.Cd.Desc=="Rape, Forcible "|Crm.Cd.Desc=="	
Other Assault "|Crm.Cd.Desc==" 	Burglary From Vehicle"|Crm.Cd.Desc=="Embezzlement, Grand Theft ($950.01 & Over) "|Crm.Cd.Desc==" 	Theft Of Identity")
```

#Use datatable to show the first 100 cases
```{r, echo=FALSE,warning=FALSE}
sample=clean_total[1:100, ]
datatable(sample, options = list(pageLength = 5,scrollX='400px'))
```

# Visualize the crime
## Preprocessing
```{r, warning=FALSE, message = FALSE}
library(tidyverse)
library("lubridate")

clean_total <- clean_total %>%
  mutate(Date = as.Date(Date.Rptd, "%m/%d/%Y %H:%M:%S")) %>%
  mutate(Location = str_squish(LOCATION))

names(clean_total)[names(clean_total) == 'Crm.Cd.Desc'] <- "Category"
clean_total$Year <- format(clean_total$Date, format="%Y")
```

# Map

## Crime over space

To visualize the data, we created an interactive map using the booklet to help us visualize the distribution of crime incidents in Los Angeles. The map contains 43,831 total crime data. The pop-up provides information on the category, date, address, victim's age, victim's gender, longitude and latitude. The purpose of this interactive map is to help us display more detailed information based on location.

```{r, echo=FALSE,warning=FALSE}
library(leaflet)

clean_total <- clean_total[clean_total$LON != 0, ] # display the first 10,000 rows
clean_total$popup <- paste("<br>", "<b>Category: </b>", clean_total$Category,
                    "<br>", "<b>Date: </b>", clean_total$Date,
                    "<br>", "<b>Address: </b>", clean_total$Location,
                    "<br>", "<b>Vict Age: </b>", clean_total$Vict.Age,
                    "<br>", "<b>Vict Sex: </b>", clean_total$Vict.Sex,
                    "<br>", "<b>Longitude: </b>", clean_total$LON,
                    "<br>", "<b>Latitude: </b>", clean_total$LAT)

leaflet(clean_total, width = "100%") %>% addTiles() %>%
  addTiles(group = "OSM (default)") %>%
  addProviderTiles(provider = "Esri.WorldStreetMap",group = "World StreetMap") %>%
  addProviderTiles(provider = "Esri.WorldImagery",group = "World Imagery") %>%
  # addProviderTiles(provider = "NASAGIBS.ViirsEarthAtNight2012",group = "Nighttime Imagery") %>%
  addMarkers(lng = ~LON, lat = ~LAT, popup = clean_total$popup, clusterOptions = markerClusterOptions()) %>%
  addLayersControl(
    baseGroups = c("OSM (default)","World StreetMap", "World Imagery"),
    options = layersControlOptions(collapsed = FALSE)
  )

```

# Crime over time
### Overall trend

The crime graph over time shows us the general trend in the number of daily crimes in Los Angeles. Based on the map, we can see that as of July 2021, Los Angeles has an annual daily crime count of around 60. In the first few months of the pandemic, crime numbers fell sharply compared to the end of 2021. Specifically, the daily crime rate dropped by nearly 78%.

```{r, echo=FALSE,warning=FALSE, message = FALSE}
library(dplyr)

df_crime_daily <- clean_total %>%
  mutate(Date = as.Date(Date, "%m/%d/%Y")) %>%
  group_by(Date) %>%
  summarize(count = n()) %>%
  arrange(Date)

library(ggplot2)
library(scales)
plot <- ggplot(df_crime_daily, aes(x = Date, y = count)) +
  geom_line(color = "#F2CA27", size = 0.1) +
  geom_smooth(color = "#1A1A1A") +
  # fte_theme() +
  scale_x_date(breaks = date_breaks("1 year"), labels = date_format("%Y")) +
  labs(x = "Date of Crime", y = "Number of Crimes", title = "Daily Crimes in Los Angeles from 2020 – Present")
plot
```

### In category

Here, I break down the crime categories into battery, burglary, intimate partner, robbery, and criminal threat. As the plot shows, we conclude that the number of burglaries and robberies has decreased significantly due to the pandemic, but the number of intimate partner crimes has increased dramatically.

```{r, echo=FALSE,warning=FALSE,message = FALSE}
library(tidytext)
library(janeaustenr)
library(dplyr)


df_crime_cat <- clean_total %>% 
  dplyr::group_by(Date,Category)%>%
  dplyr::summarise(Count = n())

library(plotly)

df_crime_cat %>% 
   ungroup() %>%
   plot_ly(x=~Date,y = ~Count, 
           color = ~ Category,
           type = "scatter",
           mode = "lines+ markers",
           opacity = 0.8
)

```


# Aggregte data

```{r, echo=FALSE,warning=FALSE, message = FALSE}
df=read.csv("/Users/kimberlyshan/Desktop/clean_total.csv")
```

### Summarize the data by crime category, and calculate the percentage of each crime.
```{r, echo=FALSE,warning=FALSE, message = FALSE}
library(DT)
library(stringr)
Sys.setlocale("LC_TIME", "C")
df_category <- sort(table(df$Crm.Cd.Desc),decreasing = TRUE)
df_category <- data.frame(df_category)
colnames(df_category) <- c("Category", "Frequency")
df_category$Percentage <- df_category$Frequency / sum(df_category$Frequency)
datatable(df_category, options = list(scrollX='400px'))
```

These are the top six categories of crime. And assaults are classified into two types: simple assault and aggravated assault. A simple assault is defined as any purposeful conduct that puts another person in reasonable fear of being battered. Aggravated assault includes more serious crimes, such as assault with the intent to cause significant bodily harm or assault with a lethal weapon, such as a firearm.


### Create a bar plot based on the crime category.
```{r, echo=FALSE,warning=FALSE}
library(ggplot2)
library(ggrepel)
bp <- ggplot(df_category, aes(x=Category, y=Frequency, fill=Category)) + 
  geom_bar(stat="identity") + 
  theme(axis.text.x=element_blank()) + 
  geom_text_repel(data=df_category, aes(label=Category), size=2)
bp
```

We can clearly observe that simple assault from intimate partner has the highest frequency number of nearly 15,000 and counts for 33% of the whole crime types. And its aggravated assault has the third highest frequency. Intimate partner violence is defined as abuse or aggressiveness in a romantic relationship. And battery has the second highest number of over 12,000 and counts for 27%. In criminal law, this is a physical act that results in damaging or offensive contact with another person without that person's agreement.


And we also create a pie chart to show percentage of each categories more straightforward.
```{r, echo=FALSE,warning=FALSE}
bp <- ggplot(df_category, aes(x="", y=Percentage, fill=Category)) +
  geom_bar(stat="identity") + 
  coord_polar("y") 
bp
```

Through these statistics, we can conclude that assault from intimate partners has the highest frequency in Los Angeles during 2020 to 2022. Many risk factors can lead to intimate partner violence, such as people who have a low self-esteem, low education or income, and unstable emotions. For relationship factors, unhealthy family relationships, strong dominance and control by the partner, etc. And people who are in the communities with easy access to drugs and alcohol are more likely to be involved in this crime. 


## Temporal Trends
### Robbery Over Time

We wanted to understand a trend in robbery in LA over the two years, as it also represents a high percentage of the overall crime types.
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_robbery <- df[df$Crm.Cd.Desc == "Robbery",]

df_robbery_daily <- df_robbery %>%
  mutate(Date = as.Date(Date.Rptd, "%m/%d/%Y %H:%M:%S")) %>%
  group_by(Date) %>%
  summarize(count = n()) %>%
  arrange(Date)

library(ggplot2)
library(scales)

plot <- ggplot(df_robbery_daily, aes(x = Date, y = count)) +
  geom_line(color = "#F2CA27", size = 0.1) +
  geom_smooth(color = "#1A1A1A") +
  # fte_theme() +
  scale_x_date(breaks = date_breaks("1 year"), labels = date_format("%Y")) +
  labs(x = "Date of Robbery", y = "Number of Robberys", title = "Daily Robberys From 2020 to 2022")
plot
```
This graph shows a downward trend of daily robbery from 2020 to present, from roughly 7 cases down to 4 cases per day. And the numbers start to decrease obviously starting from 2022.

### Robbery Time Heatmap
We want to know the most frequent days in a week of robbery.So we aggregate counts of robbery by Day-of-Week and Time to create a heat map.
```{r, echo=FALSE,warning=FALSE, message = FALSE}
library(lubridate)

get_hour <- function(x) {
  return (as.numeric((str_split(str_split(x, " ", simplify = T)[,2], ":", simplify = T)[,1])))
}

get_weekdays <- function(x){
  return (weekdays(as.Date(x, format = "%m/%d/%Y %H:%M:%S")))
}

df_robbery_time <- df_robbery %>%
  mutate(Hour = sapply(Date.Rptd, get_hour)) %>%
  mutate(DayOfWeek = sapply(Date.Rptd, get_weekdays)) %>%
  group_by(DayOfWeek, Hour) %>%
  summarize(count = n())
# df_theft_time %>% head(10)
datatable(df_robbery_time, options = list(scrollX='400px'))
```
From the table, we can see that Friday, Monday, and Saturdays are the top three frequent days in a week of robbery. And Monday has the highest counts of 675 over the time.

Create a time heatmap.
```{r, echo=FALSE,warning=FALSE}
plot <- ggplot(df_robbery_time, aes(x = Hour, y = DayOfWeek, fill = count)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.6), legend.title = element_blank(), legend.position="top", legend.direction="horizontal", legend.key.width=unit(2, "cm"), legend.key.height=unit(0.25, "cm"), legend.margin=unit(-0.5,"cm"), panel.margin=element_blank()) +
  labs(x = "Hour of Roberry (Local Time)", y = "Day of Week of Roberry", title = "Number of Roberry from 2020 to 2022, by Time of Roberry") +
  scale_fill_gradient(low = "white", high = "#27AE60", labels = comma)
plot
```
From the graph, between 11:50 to 12:50AM, robbery happens the most on Mondays, more than 660 cases. Comparing to Mondays, Fridays have the least numbers of rubbery during that time.

### Burglary Over Time
Then we want to know if there is any difference or common between robbery and burglary, so we create a chart of crimes (Burglary) over time.
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_burglary <- df[df$Crm.Cd.Desc == "Burglary",]

df_burglary_daily <- df_burglary %>%
  mutate(Date = as.Date(Date.Rptd, "%m/%d/%Y %H:%M:%S")) %>%
  group_by(Date) %>%
  summarize(count = n()) %>%
  arrange(Date)

library(ggplot2)
library(scales)

plot <- ggplot(df_burglary_daily, aes(x = Date, y = count)) +
  geom_line(color = "#F2CA27", size = 0.1) +
  geom_smooth(color = "#1A1A1A") +
  # fte_theme() +
  scale_x_date(breaks = date_breaks("1 year"), labels = date_format("%Y")) +
  labs(x = "Date of Burglary", y = "Number of Burglarys", title = "Daily Burglarys From 2020 to 2022")
plot
```

Comparing to robbery, the trend line of burglary is relatively smooth with an average number of 4 per day, and it begins to slightly decrease in 2022.

We also create a time heatmap of burglary.
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_burglary_time <- df_burglary %>%
  mutate(Hour = sapply(Date.Rptd, get_hour)) %>%
  mutate(DayOfWeek = sapply(Date.Rptd, get_weekdays)) %>%
  group_by(DayOfWeek, Hour) %>%
  summarize(count = n())

plot <- ggplot(df_burglary_time, aes(x = Hour, y = DayOfWeek, fill = count)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.6), legend.title = element_blank(), legend.position="top", legend.direction="horizontal", legend.key.width=unit(2, "cm"), legend.key.height=unit(0.25, "cm"), legend.margin=unit(-0.5,"cm"), panel.margin=element_blank()) +
  labs(x = "Hour of Burglary (Local Time)", y = "Day of Week of Burglary", title = "Number of Burglary from 2020 to 2022, by Time of Burglary") +
  scale_fill_gradient(low = "white", high = "#27AE60", labels = comma)
plot
```

From the map, we can see Mondays still are the most frequent happening days between 11:50 to 12:50AM, but now Saturdays have the least numbers of burglary. Burglary is when someone illegally enters a building in order to conduct a crime while inside; robbery is when someone steals anything of value directly from another person using force or terror. So we assume the reason that lead to Monday being the highest frequency of robberies is people are out for lunch, many people are on the streets, which give thieves more opportunities. As for burglary, most people are not home during the weekdays, so thieves will likely take advantage and break into the houses, and vice versa.

# Correlation analysis

```{r, echo=FALSE,warning=FALSE, message = FALSE}
library(magrittr)
library(dplyr)
library(DT)
library(ggplot2)
library(gridExtra)
library(grid)
```

```{r, echo=FALSE,warning=FALSE}
df=read.csv("/Users/kimberlyshan/Desktop/clean_total.csv")
# Convert date formate
df$date <- substr(df$Date.Rptd,1,10)
betterDates <- as.Date(df$date, "%m/%d/%y")

df$DayOfWeek <- weekdays(betterDates)

```

### Crime By Category
```{r, echo=FALSE,warning=FALSE}
df_arrest <- df %>% filter(grepl("Arrest", Status.Desc))
df_top_crimes <- df_arrest %>%
  group_by(Crm.Cd.Desc) %>% 
  summarize(count = n()) %>%
  arrange(desc(count))

datatable(df_top_crimes, options = list(pageLength = 10,scrollX='400px'))
```

```{r, echo=FALSE,warning=FALSE}
# Define a function that will conver Time.Occ to two-digit time indicator
# For instance, 1245 will return 12, 25 will return 0, and 345 will return 3
ToTime <- function(x){
  if ((floor(log10(x)) + 1)<=2){
    result <- "0"
  } else if (floor(log10(x)) + 1==3){
    result <- substr(x, 1, 1)
  } else {
    result <- substr(x, 1, 2)
  }
  return (result)
}

```


### By Category
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_arrest_time_crime <- df_arrest %>%
  filter(Crm.Cd.Desc %in% df_top_crimes$Crm.Cd.Desc[2:19]) %>%
  mutate(Hour = sapply(TIME.OCC, ToTime)) %>%
  group_by(Crm.Cd.Desc, DayOfWeek, Hour) %>% 
  summarize(count = n())

dow_format <- c("Sunday","Monday","Tuesday","Wednesday","Thursday","Friday","Saturday")
hour_format <- c(paste(c(12,1:11),"AM"), paste(c(12,1:11),"PM"))

df_arrest_time_crime$DayOfWeek <- factor(df_arrest_time_crime$DayOfWeek, level = rev(dow_format))
df_arrest_time_crime$Hour <- factor(df_arrest_time_crime$Hour, level = 0:23, label = hour_format)

datatable(df_arrest_time_crime, options = list(pageLength = 10, scrollX='400px'))
```



```{r, echo=FALSE,warning=FALSE}

plot <- ggplot(df_arrest_time_crime, aes(x = Hour, y = DayOfWeek, fill = count)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text=element_text(size=8), axis.text.x = element_text(angle = 90, vjust = 0.6, size = 8)) +
  labs(x = "Hour of Arrest (Local Time)", y = "Day of Week of Arrest", title = "Number of Police Arrests in Los Angeles from 2020 – Present, by Category and Time of Arrest") +
  scale_fill_gradient(low = "white", high = "#2980B9") +
  facet_wrap(~ Crm.Cd.Desc, nrow = 3)
plot

```
## By Category (Normalized)
```{r, echo=FALSE,warning=FALSE}
df_arrest_time_crime <- df_arrest_time_crime %>%
  group_by(Crm.Cd.Desc) %>%
  mutate(norm = count/sum(count))

datatable(df_arrest_time_crime, options = list(pageLength = 10,scrollX='400px'))

plot <- ggplot(df_arrest_time_crime, aes(x = Hour, y = DayOfWeek, fill = norm)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.6, size = 8)) +
  labs(x = "Hour of Arrest (Local Time)", y = "Day of Week of Arrest", title = "Police Arrests in Los Angeles from 2020 by Time of Arrest, Normalized by Type of Crime") +
  scale_fill_gradient(low = "white", high = "#2980B9") +
  facet_wrap(~ Crm.Cd.Desc, nrow = 3)
plot


```
## By Police District
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_arrest_time_district <- df_arrest %>%
  mutate(Hour = sapply(TIME.OCC, ToTime)) %>%
  group_by(AREA.NAME, DayOfWeek, Hour) %>% 
  summarize(count = n()) %>%
  group_by(AREA.NAME) %>%
  mutate(norm = count/sum(count))

df_arrest_time_district$DayOfWeek <- factor(df_arrest_time_district$DayOfWeek, level = rev(dow_format))
df_arrest_time_district$Hour <- factor(df_arrest_time_district$Hour, level = 0:23, label = hour_format)

datatable(df_arrest_time_district, options = list(pageLength = 10,scrollX='400px'))
```
```{r,echo=FALSE, warning=FALSE}
plot <- ggplot(df_arrest_time_district, aes(x = Hour, y = DayOfWeek, fill = norm)) +
  geom_tile() +
  theme(axis.text=element_text(size=5),axis.text.x = element_text(angle = 90, vjust = 0.6, size =4)) +
  labs(x = "Hour of Arrest (Local Time)", y = "Day of Week of Arrest", title = "Police Arrests in Los Angeles from 2020 by Time of Arrest, Normalized by Station") +
  scale_fill_gradient(low = "white", high = "#8E44AD") +
  facet_wrap(~ AREA.NAME, nrow=5,shrink=FALSE) 
plot



```

### By Month
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_arrest_time_month <- df_arrest %>%
  mutate(Month = format(as.Date(date, "%m/%d/%Y"), "%B"), Hour = sapply(TIME.OCC, ToTime)) %>%
  group_by(Month, DayOfWeek, Hour) %>% 
  summarize(count = n()) %>%
  group_by(Month) %>%
  mutate(norm = count/sum(count))

df_arrest_time_month$DayOfWeek <- factor(df_arrest_time_month$DayOfWeek, level = rev(dow_format))
df_arrest_time_month$Hour <- factor(df_arrest_time_month$Hour, level = 0:23, label = hour_format)

# Set order of month facets by chronological order instead of alphabetical
df_arrest_time_month$Month <- factor(df_arrest_time_month$Month,
                                     level = c("January","February","March","April","May","June","July","August","September","October","November","December"))

plot <- ggplot(df_arrest_time_month, aes(x = Hour, y = DayOfWeek, fill = norm)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.6, size = 4)) +
  labs(x = "Hour of Arrest (Local Time)", y = "Day of Week of Arrest", title = "Police Arrests in Los Angeles from 2020 by Time of Arrest, Normalized by Month") +
  scale_fill_gradient(low = "white", high = "#E74C3C") +
  facet_wrap(~ Month, nrow = 4)
plot
```

### By Year
```{r, echo=FALSE,warning=FALSE, message = FALSE}
df_arrest_time_year <- df_arrest %>%
  mutate(Year = format(as.Date(date(), "%m/%d/%Y"), "%Y"), Hour = sapply(TIME.OCC, ToTime)) %>%
  group_by(Year, DayOfWeek, Hour) %>% 
  summarize(count = n()) %>%
  group_by(Year) %>%
  mutate(norm = count/sum(count))

df_arrest_time_year$DayOfWeek <- factor(df_arrest_time_year$DayOfWeek, level = rev(dow_format))
df_arrest_time_year$Hour <- factor(df_arrest_time_year$Hour, level = 0:23, label = hour_format)

plot <- ggplot(df_arrest_time_year, aes(x = Hour, y = DayOfWeek, fill = norm)) +
  geom_tile() +
  # fte_theme() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.6, size = 4)) +
  labs(x = "Hour of Arrest (Local Time)", y = "Day of Week of Arrest", title = "Police Arrests Los Angeles from 2020 by Time of Arrest, Normalized by Year") +
  scale_fill_gradient(low = "white", high = "#E67E22") +
  facet_wrap(~ Year, nrow = 6)
plot
```