Introduction:

The CDC’s “PLACES: Local Data for Better Health” provides access to detailed local health data. It helps users explore health indicators by county, city, and census areas to improve public health efforts. The platform offers interactive maps, data tools, and resources for understanding local health measures based on CDC and U.S. Census data. For more details, visit the PLACES website.

This dataset is adapted from 500 Cities: Local Data for Better Health, 2017 release.

Project Prompt:

For this project, you will work with a cleaned dataset and conduct an analysis using GIS techniques.

  1. Start by filtering the dataset further to create a subset containing no more than 900 observations. Choose a specific subset based on a meaningful criterion related to your analysis.

  2. Create a plot that visualizes an aspect of your subsetted dataset. This could be a histogram, scatter plot, or line chart, depending on the nature of your data.

  3. Generate a basic GIS map that represents the geographic distribution of your subsetted data points. Ensure that the map clearly conveys relevant spatial patterns.

  4. Refine your GIS map by adding interactive elements, such as a tooltip that displays information when users click on a data point.

  5. Write a paragraph summarizing your visualizations. Explain what your plot and map reveal about your subsetted dataset. Discuss any trends, patterns, or insights gained from your analysis.

This project will help you practice data filtering, visualization, and GIS mapping techniques, reinforcing concepts from the Japan earthquakes tutorial.

Let’s start:

Load the libraries and set the working directory

library(tidyverse)
library(tidyr)
library(leaflet)
library(sf)
library(knitr)

setwd("C:/Users/hanle/Desktop/Project 2")

cities500 <- read_csv("500CitiesLocalHealthIndicators.cdc.csv")
data(cities500)

Cleaning the data set:

1. The GeoLocation variable has (lat, long) format. We need to split GeoLocation (lat, long) into two columns: lat and long.

To do so, we will remove the parentheses from a column, then split it into separate latitude and longitude columns.

latlong <- cities500|>
  mutate(GeoLocation = str_replace_all(GeoLocation, "[()]", ""))|>
  separate(GeoLocation,
           into = c("lat", "long"),
           sep = ",",
           convert = TRUE)
head(latlong)
## # A tibble: 6 × 25
##    Year StateAbbr StateDesc  CityName  GeographicLevel DataSource Category      
##   <dbl> <chr>     <chr>      <chr>     <chr>           <chr>      <chr>         
## 1  2017 CA        California Hawthorne Census Tract    BRFSS      Health Outcom…
## 2  2017 CA        California Hawthorne City            BRFSS      Unhealthy Beh…
## 3  2017 CA        California Hayward   City            BRFSS      Health Outcom…
## 4  2017 CA        California Hayward   City            BRFSS      Unhealthy Beh…
## 5  2017 CA        California Hemet     City            BRFSS      Prevention    
## 6  2017 CA        California Indio     Census Tract    BRFSS      Health Outcom…
## # ℹ 18 more variables: UniqueID <chr>, Measure <chr>, Data_Value_Unit <chr>,
## #   DataValueTypeID <chr>, Data_Value_Type <chr>, Data_Value <dbl>,
## #   Low_Confidence_Limit <dbl>, High_Confidence_Limit <dbl>,
## #   Data_Value_Footnote_Symbol <chr>, Data_Value_Footnote <chr>,
## #   PopulationCount <dbl>, lat <dbl>, long <dbl>, CategoryID <chr>,
## #   MeasureId <chr>, CityFIPS <dbl>, TractFIPS <dbl>, Short_Question_Text <chr>

str_replace_all(GeoLocation, "[()]", "") removes any parentheses from the GeoLocation column.

separate(GeoLocation, into = c("lat", "long"), sep = ",", convert = TRUE)

The separate() function splits the GeoLocation column into two new columns:

“lat” (latitude) “long” (longitude)

sep = "," specifies that the values are separated by a comma.

convert = TRUE automatically converts the new columns into appropriate data types (numeric in this case).

2. Filter the dataset: Remove the StateDesc that includes the United Sates, select Prevention as the category (of interest), filter for only measuring crude prevalence and select only 2017.

latlong_clean <- latlong |>
  filter(StateDesc != "United States") |>
  filter(Category == "Prevention") |>
  filter(Data_Value_Type == "Crude prevalence") |>
  filter(Year == 2017)
head(latlong_clean)
## # A tibble: 6 × 25
##    Year StateAbbr StateDesc  CityName   GeographicLevel DataSource Category  
##   <dbl> <chr>     <chr>      <chr>      <chr>           <chr>      <chr>     
## 1  2017 AL        Alabama    Montgomery City            BRFSS      Prevention
## 2  2017 CA        California Concord    City            BRFSS      Prevention
## 3  2017 CA        California Concord    City            BRFSS      Prevention
## 4  2017 CA        California Fontana    City            BRFSS      Prevention
## 5  2017 CA        California Richmond   Census Tract    BRFSS      Prevention
## 6  2017 FL        Florida    Davie      Census Tract    BRFSS      Prevention
## # ℹ 18 more variables: UniqueID <chr>, Measure <chr>, Data_Value_Unit <chr>,
## #   DataValueTypeID <chr>, Data_Value_Type <chr>, Data_Value <dbl>,
## #   Low_Confidence_Limit <dbl>, High_Confidence_Limit <dbl>,
## #   Data_Value_Footnote_Symbol <chr>, Data_Value_Footnote <chr>,
## #   PopulationCount <dbl>, lat <dbl>, long <dbl>, CategoryID <chr>,
## #   MeasureId <chr>, CityFIPS <dbl>, TractFIPS <dbl>, Short_Question_Text <chr>

What variables are included? (can any of them be removed?)

names(latlong_clean)
##  [1] "Year"                       "StateAbbr"                 
##  [3] "StateDesc"                  "CityName"                  
##  [5] "GeographicLevel"            "DataSource"                
##  [7] "Category"                   "UniqueID"                  
##  [9] "Measure"                    "Data_Value_Unit"           
## [11] "DataValueTypeID"            "Data_Value_Type"           
## [13] "Data_Value"                 "Low_Confidence_Limit"      
## [15] "High_Confidence_Limit"      "Data_Value_Footnote_Symbol"
## [17] "Data_Value_Footnote"        "PopulationCount"           
## [19] "lat"                        "long"                      
## [21] "CategoryID"                 "MeasureId"                 
## [23] "CityFIPS"                   "TractFIPS"                 
## [25] "Short_Question_Text"

Remove the variables that will not be used in the assignment

prevention <- latlong_clean |>
  select(-DataSource,-Data_Value_Unit, -DataValueTypeID, -Low_Confidence_Limit, -High_Confidence_Limit, -Data_Value_Footnote_Symbol, -Data_Value_Footnote)
head(prevention)
## # A tibble: 6 × 18
##    Year StateAbbr StateDesc  CityName  GeographicLevel Category UniqueID Measure
##   <dbl> <chr>     <chr>      <chr>     <chr>           <chr>    <chr>    <chr>  
## 1  2017 AL        Alabama    Montgome… City            Prevent… 151000   Choles…
## 2  2017 CA        California Concord   City            Prevent… 616000   Visits…
## 3  2017 CA        California Concord   City            Prevent… 616000   Choles…
## 4  2017 CA        California Fontana   City            Prevent… 624680   Visits…
## 5  2017 CA        California Richmond  Census Tract    Prevent… 0660620… Choles…
## 6  2017 FL        Florida    Davie     Census Tract    Prevent… 1216475… Choles…
## # ℹ 10 more variables: Data_Value_Type <chr>, Data_Value <dbl>,
## #   PopulationCount <dbl>, lat <dbl>, long <dbl>, CategoryID <chr>,
## #   MeasureId <chr>, CityFIPS <dbl>, TractFIPS <dbl>, Short_Question_Text <chr>
md <- prevention |>
  filter(StateAbbr=="MD")
dim(md)
## [1] 804  18

Your Work

  1. MD has 804 observations. For this project, start by filtering the dataset further to create a subset containing no more than 900 observations. Choose a specific subset based on a meaningful criterion related to your analysis.
de_chol <- prevention |>
  filter(StateAbbr=="DE") |>
  filter(Measure == "Cholesterol screening among adults aged >=18 Years")

dim(de_chol)
## [1] 25 18
  1. Create a plot that visualizes an aspect of your subsetted dataset. This could be a histogram, scatter plot, or line chart, depending on the nature of your data.
ggplot(de_chol, aes(x = long, 
                    y = lat,
                    color = Data_Value)) +
  geom_point(aes(size = PopulationCount)) +
  scale_color_gradient(high = "#8fceff",
                       low = "#000000") +
  theme_bw() +
  labs(title = "Map of Cholesterol Screening Rate with Population as Bubble Size",
       x = "Longitude", 
       y = "Latitude",
       caption = "Source: CDC's PLACES: Local Data for Better Health")

  1. Generate a basic GIS map that represents the geographic distribution of your subsetted data points. Ensure that the map clearly conveys relevant spatial patterns.
leaflet() |>
  setView(lng = -75.5484, lat = 39.7447, zoom = 13) |>
  addProviderTiles("Esri.NatGeoWorldMap") |>
  addCircles(
    data = de_chol,
    radius = de_chol$PopulationCount/80,
    color = "#000000",
    fillColor = "#740fba",
    fillOpacity = 0.25
)
## Assuming "long" and "lat" are longitude and latitude, respectively
  1. Refine your GIS map by adding interactive elements, such as a tooltip that displays information when users click on a data point.
popupdel <- paste("Coordinates:", de_chol$lat,de_chol$long,
                   "<br>Cholesterol Screening Rate:", de_chol$Data_Value, "%",
                   "<br>Population:", de_chol$PopulationCount)
leaflet() |>
  setView(lng = -75.5484, lat = 39.7447, zoom = 13) |>
  addProviderTiles("Esri.NatGeoWorldMap") |>
  addCircles(
    data = de_chol,
    radius = de_chol$PopulationCount/80,
    color = "#000000",
    fillColor = "#740fba",
    fillOpacity = 0.25,
    popup = popupdel) 
## Assuming "long" and "lat" are longitude and latitude, respectively
  1. Write a paragraph summarizing your visualizations. Explain what your plot and map reveal about your subsetted dataset. Discuss any trends, patterns, or insights gained from your analysis.

The visualizations I created focus on the relationship between population size and cholesterol screening rates across census tracts in specifically Wilmington, Delaware.

In the scatter plot, each circle represents a census tract. The x-axis and y-axis are based off the latitude and longitude of each observation. This makes the scatter plot mimic how the points would look on a real map. The size of the circle represent the population of each census tract; the bigger the size, the bigger the population. Additionally, the lighter a color of a point is, the higher the cholesterol screening rate of the census tract. The scatterplot shows that there is not a big correlation between population and cholesterol screening rate, as both populated and less populated census tracts vary in rates.

The Leaflet map provides a more detailed and interactive view of what was displayed in the scatter plot. It reveals that there was one census tract with a much higher population than the rest. The tooltips show that in that specific census tract, the population was around 71k; whereas when you click on the other census tracts, many are less than 10k. However, the tooltips also show that screening rates do not rely on population size. For example, the census tract with the biggest population has a cholesterol screening rate of 78.1%, a census tract with a population of 2303 has a screening rate of 82.6%, and a census tract with a population of 3275 has a rate of 64.9%. This suggests that population size does not strongly affect cholesterol screening rates in Wilmington, Delaware.