knitr::opts_chunk$set( echo = TRUE, message = FALSE, warning = FALSE, collapse = TRUE, comment = "#>", fig.align = "center", fig.retina = 2)
pdxTrees
is a data package composed of information on inventoried trees in Portland, OR. There are two datasets that can be accessed with this package:
get_pdxTrees_parks()
pulls in data on up to 25,534 trees from 174 Portland parks.
get_pdxTrees_streets()
pulls in data on up to 218,602 trees located on Portland's streets. A street tree is loosely defined as a tree generally in the public right-of-way, usually between the sidewalk and the street.
The street trees are categorized by one of the 96 Portland neighborhoods and the park trees are categorized by the public parks in which they grow.
Here are some examples of the different ways pdxTrees
can be used in an educational setting!
# First make sure you have the package downloaded! # devtools::install_github("mcconvil/pdxTrees") # Loading the required libraries library(pdxTrees) library(ggplot2) library(dplyr) library(forcats)
First we have to grab the data. To do this we use the get_pdxTrees_parks()
and get_pdxTrees_streets()
functions. In this vignette, we only explore the parks dataset.
# Leaving the argument field blank pulls data for all of the parks! pdxTrees_parks <- get_pdxTrees_parks()
ggplot2
# A histogram of the inventory date pdxTrees_parks %>% count(Inventory_Date) %>% # Setting the aesthetics ggplot(aes(x = Inventory_Date)) + # Specifying a histogram and picking color! geom_histogram(bins = 50, fill = "darkgreen", color = "black") + labs( x = "Inventory Date", y = "Count", title= "When was pdxTrees_parks Inventoried?") + # Adding a theme theme_minimal() + theme(plot.title = element_text(hjust = 0.5))
Using ggplot2
we can create a histogram of the pdxTrees_parks
inventory dates. The trees were inventoried from 2017 to 2019 with the majority of the trees inventoried in the summer months, when the weather is nice in Portland.
This graph is just one of example of how pdxTrees
can be used to create data visualizations. With a healthy mix of categorical and quantitative variables in both datasets, you can make scatterplots, bar graphs, density plots, etc. For more advanced visualizations, you can add animation with gganimate
or create an interactive map with leaflet
.
leaflet
The following code creates an interactive map with leaflet
and the pdxTrees_parks
data. It showcases:
```{R leaflet packages}
library(leaflet) library(leaflet.extras)
```r # Making the leaf popup icon greenLeaflittle <- makeIcon( iconUrl = "http://leafletjs.com/examples/custom-icons/leaf-green.png", iconWidth = 10, iconHeight = 20, iconAnchorX = 10, iconAnchorY = 10, shadowUrl = "http://leafletjs.com/examples/custom-icons/leaf-shadow.png", shadowWidth = 10, shadowHeight = 15, shadowAnchorX = 5, shadowAnchorY = 5 ) # Pulling the data for Berkely Park berkeley_prk <- get_pdxTrees_parks(park = "Berkeley Park") # Creating the popup label labels <- paste("</b>", "Common Name:", berkeley_prk$Common_Name, "</b></br>", "Factoid: ", berkeley_prk$Species_Factoid) # Creating the map leaflet() %>% # Setting the lng and lat to be in the general area of Berekely Park setView(lng = -122.6239, lat = 45.4726, zoom = 17) %>% # Setting the background tiles addProviderTiles(providers$Esri.WorldTopoMap) %>% # Adding the leaf markers with the popup data on top of the circles markers addMarkers( ~Longitude, ~Latitude, data = berkeley_prk, icon = greenLeaflittle, popup = ~labels) %>% # Adding the mini map at the bottom right corner addMiniMap()
gganimate
library(gganimate)
Before you animate a graph with gganimate
you have to create and save a graph with ggplot2
.
# Refactoring the categorical mature_size variable berkeley_prk <- berkeley_prk %>% mutate(mature_size = fct_relevel(Mature_Size, "S", "M", "L")) # First creating the graph using ggplot and saving it! berkeley_graph <- berkeley_prk %>% # Piping in the data ggplot(aes(x = Tree_Height, y = Pollution_Removal_value, color = Mature_Size)) + # Creating the scatterplot geom_point(size = 2, alpha = 0.5) + theme_minimal() + # Adding the labels labs(title = "Pollution Removal Value of Berkeley Park Trees", x = "Tree Height", y = "Pollution Removal Value ($'s annually)", color = "Mature Size") + # Adding a color palette scale_color_brewer(type = "seq", palette = "Set1") + # Customizing the title font theme(plot.title = element_text(hjust = 0.5, size = 8, face = "bold"), axis.title.x = element_text(size = 6), axis.text = element_text(size = 4), axis.title.y = element_text(size = 6), legend.title = element_text(size = 6), legend.text = element_text(size= 4))
Now we can add animation!
```{R, out.width = "90%"}
berkeley_graph +
# Choosing which variable we want to annimate
transition_states(states = Mature_Size,
# How long each point stays before fading away
transition_length = 10,
# Time the transition takes
state_length = 8) +
# Animation for the points entering
enter_grow() +
# Animation for the points exiting
exit_shrink()
Unsurprisingly it seems that large trees have the highest degree of pollution removal. However, there is a lot of overlap between the categories that could likely be explained by other key variables, like `Species`. This is a great opportunity to practice multivariate thinking! ## A Linear Regression Aside from visualizations, the data can also be used to build linear regression models, create confidence intervals, and perform many other forms of statistical inference. Let's look at the relationship between `Tree_Height` and `Pollution_Removal_value`. ```{R linear regression graph, warning = FALSE, fig.width = 6, fig.height = 4} # Visualizing the relationship between the two variables. ggplot(pdxTrees_parks, aes(x = Tree_Height, y = Pollution_Removal_value)) + # Creating a scatter plot geom_point(alpha = 0.05) + # Adding the line of best fit stat_smooth(method = lm, se = FALSE) + theme_minimal() + labs(x = "Tree Height", y = "Pollution Removal Value ($)")
```{R linear regression, warning = FALSE}
library(moderndive)
mod <- lm(Pollution_Removal_value ~ Tree_Height, data = pdxTrees_parks)
get_regression_table(mod) ```
Looking at the graph and the slope coefficient from the regression table, it does seem like tree height positively correlates with pollution removal. In particular, with a $\hat\beta_1$ of $0.104$, we'd estimate that the Pollution_Removal_value
increases by $0.104$ with each additional foot of Tree_Height
.
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.