Project 2

For this project I am going to be cleaning and looking through a dataset that focuses on Accidental drug related deaths from 2012-2022 in the state of Connecticut. I am going to foucs on the amount of deaths for each drug and also how that death happened so we can see which drugs have the mose affect on people in Connecticut. We are also going to take a look at the age range of these death to see what age group certain drugs are affecting people.

Loading in the needed libraries for this project. I am also setting my dedicatory

library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.4     ✔ readr     2.1.5
✔ forcats   1.0.0     ✔ stringr   1.5.1
✔ ggplot2   3.4.4     ✔ tibble    3.2.1
✔ lubridate 1.9.3     ✔ tidyr     1.3.0
✔ purrr     1.0.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(highcharter)

Registered S3 method overwritten by 'quantmod':
  method            from
  as.zoo.data.frame zoo 

setwd("C:/Users/chris/Downloads/Data Science R")

Now I am loading in the dataset and changing the name to something easier. I am also taking a quick look at some of the data

# Load the data
drug_deaths <- read_csv("Accidental_Drug_Related_Deaths_2012-2022_20240410.csv")

Rows: 10654 Columns: 48
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (47): Date, Date Type, Sex, Race, Ethnicity, Residence City, Residence C...
dbl  (1): Age

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

head(drug_deaths)

# A tibble: 6 × 48
  Date       `Date Type`     Age Sex    Race  Ethnicity `Residence City`
  <chr>      <chr>         <dbl> <chr>  <chr> <chr>     <chr>           
1 5/29/2012  Date of death    37 Male   Black <NA>      STAMFORD        
2 6/27/2012  Date of death    37 Male   White <NA>      NORWICH         
3 3/24/2014  Date of death    28 Male   White <NA>      HEBRON          
4 12/31/2014 Date of death    26 Female White <NA>      BALTIC          
5 1/16/2016  Date of death    41 Male   White <NA>      SHELTON         
6 6/13/2017  Date reported    57 Male   White <NA>      BLANDFORD       
# ℹ 41 more variables: `Residence County` <chr>, `Residence State` <chr>,
#   `Injury City` <chr>, `Injury County` <chr>, `Injury State` <chr>,
#   `Injury Place` <chr>, `Description of Injury` <chr>, `Death City` <chr>,
#   `Death County` <chr>, `Death State` <chr>, Location <chr>,
#   `Location if Other` <chr>, `Cause of Death` <chr>, `Manner of Death` <chr>,
#   `Other Significant Conditions` <chr>, Heroin <chr>,
#   `Heroin death certificate (DC)` <chr>, Cocaine <chr>, Fentanyl <chr>, …

Now after looking at the data I am going to convert all of the Y’s into 1’s so we can have another quantitative colmun and so we can read the data better. We are going to make a different dataset that only has the amount of deaths for each drug added up. I am also going to only have the year and the top 5 drugs in a new dataset and use that to make my graphs

# Converting drug-related 'Y'/'N' to 1/0 and parsing dates

#Identify columns where the drug presence is indicated by 'Y' abd absence by 'N' or NA
drug_columns <- names(drug_deaths)[sapply(drug_deaths, function(x) any(x == 'Y', na.rm = TRUE))]

# This converts the columns from 'Y'/'N' to 0's and 1's so we can read it better
drug_deaths[drug_columns] <- drug_deaths[drug_columns] |>
  mutate_all(~ifelse(. == 'Y', 1, 0)) #If it is 'Y' then it will be converted to 1, else converts to 0

drug_deaths$Date <- mdy(drug_deaths$Date)  # Adjust based on actual date format

#Adds a new column "Year" that was extracted from the "Date" column
drug_deaths$Year <- year(drug_deaths$Date)
total_deaths_per_drug <-  colSums(drug_deaths[drug_columns], na.rm = TRUE) #Calculates the total number of deaths per drug
top_drugs <- names(sort(total_deaths_per_drug, decreasing = TRUE))[1:5] #Sorts so we can get the top 5 drugs that have the most deaths

# Summarizing deaths per year for the top 5 drugs and handling missing data
annual_drug_summary <- drug_deaths |>
  group_by(Year) |>
  summarise_at(vars(one_of(top_drugs)), sum, na.rm = TRUE) |> # Sums up the deaths for each top 5 drug by year
  ungroup() |>
  complete(Year = full_seq(Year, 1), fill = list(Heroin = 0, Fentanyl = 0, Cocaine = 0, Others1 = 0, Others2 = 0)) # Adds zeros to missing years

# Reshaping the data from wide to long format so it is easier to use ggplot
drug_plot_data <- pivot_longer(annual_drug_summary, cols = top_drugs, names_to = "Drug", values_to = "Deaths")

Warning: Using an external vector in selections was deprecated in tidyselect 1.1.0.
ℹ Please use `all_of()` or `any_of()` instead.
  # Was:
  data %>% select(top_drugs)

  # Now:
  data %>% select(all_of(top_drugs))

See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.

Now i am going to create a line graph that shows the amount of deaths for each top 5 drug through out the years

# Creating a line graph of the top 5 drugs with the most deaths
ggplot(drug_plot_data, aes(x = Year, y = Deaths, color = Drug)) +
  geom_line(size = 1.2) +
  scale_colour_viridis_d(option = "magma") +
  labs(title = "Annual Drug-Related Deaths by Drug",
       subtitle = "Top 5 Drugs",
       x = "Year",
       y = "Number of Deaths",
       caption = "Date is from data.ct.gov") +
  theme_bw() +
  theme_dark() +
  scale_color_viridis_d(begin = 0.1, end = 0.9)  # Using a viridis color scale for better visibility

Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.

Scale for colour is already present.
Adding another scale for colour, which will replace the existing scale.

Now I am going to make a high charter verision of the line graph above so it looks way better and you can see how many deaths when you hover over it

annual_melted <- pivot_longer(annual_drug_summary, cols = top_drugs, names_to = "Drug", values_to = "Deaths")

hc <- highchart() |>
  hc_title(text = "Annual Drug-Related Deaths (2012-2022") |>
  hc_subtitle(text = "Top 5 Drugs") |>
  hc_xAxis(title = list(text = "Year")) |>
  hc_yAxis(title = list(text = "Number of Deaths")) |>
  hc_legend(enabled = TRUE) |>
  hc_add_theme(hc_theme_superheroes()) |>
  hc_chart(zoomType = "x") |>
  hc_plotOptions(series = list(marker = list(enabled = TRUE)))

  


# Asked chatgpt for help bc without this it was not working. It said it was needed to add multiple series to the highcharter object. 
for(drug in top_drugs) {
  drug_data <- annual_melted |> filter(Drug == drug)
  hc <- hc |> hc_add_series(data = drug_data, type = "line", name = drug, hcaes(x = Year, y = Deaths))
}

hc

Questions

A. My topic was drug related deaths in Connecticut from 2012-2022 and I choose this topic because I wanted to see how drugs where impacting states and to see which drug was the worst and so I found a dataset that was for Connecticut and used it. There were a lot of variables included in this dataset, some are date, sex, age, race, cause of death, the types of drugs that caused deaths, where it happened, and many more. They were mostly categorical and there was only 1 quantitative but later on I made more quantitative variables. The data came from CT data (https://data.ct.gov/Health-and-Human-Services/Accidental-Drug-Related-Deaths-2012-2022/rybz-nyjw/about_data). I cleaned it up so we could see the amount of deaths per year per drug by changing some variables to quantitative.

B. I found this article that I believe also used the same sources as me. https://portal.ct.gov/-/media/dph/injury-and-violence-prevention/injuryfactsheets/2022-fact-sheet_unintentional-fatal-drug-overdoses_updated-on-6-7-2023.pdf. But this shows some different stats that I found interesting. It also shows some different graphs about the data different from mine

C. When I finished making the graph I found out that it wasn’t just fentanyl that had the highest deaths but also any opioids, even though that any opioids is pretty general, it still surprised me with how many deaths it had throughout the years. I also noticed that opioids and fentanyl around 2018 started to follow each other which could tell us that fentanyl was being laced into opioids. The first thing I wanted to make was a leaflet to show where the people were dying, from what drug, and the cause of death but as I got further into coding and cleaning, I noticed it was going to be harder than I excpted so I decided to change my appraoch and change thigns up but maybe in the future as a mini project I will do that.