Data Science with R for the Life Sciences

Spatial Data Vector

Notes

Remember to download and put into data subdirectory:

Load the following into browser window:

  • [Vector Description]

Set-up R Console:

library(ggplot2)

Introduction to Vector Data

  • Vector data includes points, lines, and polygons
  • Examples include geopolitical boundaries, the location of field observations, and roads

Diagram of different types of vector data. Points: Shows 4 points each associated with a pair of x,y values Lines: Shows two lines composed on connected sets of x,y points Polygons: Shows two polygons composed of connected sets of x,y points where the last point is connected to the first point making the polygons "closed"

  • Vector data comes in a variety of formats
  • shapefiles are are one of the most common
  • They set of multiple files with the same name, but with different extensions
  • We can see this by looking at the data in data/HARV
  • This data includes data on some field plots at that Harvard Forest NEON site we’ve been working with

  • It is stored in the plots_harv files and we can see there are four of them with different extensions

  • Work with vector data using the sf package
  • We can read this data into R using st_read
  • Let’s load in the plot data we just look at
library(sf)
plots_harv <- st_read("data/HARV/harv_plots.shp")
  • When read read the data in we see information about it including
  • The data has 5 features
  • Each feature is one object, either a point, a line, or a polygon
  • The geometry type is “POINT”, which means that the features are points
  • The data has 2 fields
  • Each field is a piece of information that is associated with each feature
  • And there is information on the minimum and maximum spatial values in the dataset
  • If we view this object we’ll see that it is a data frame with one row per vector object
  • There are three columns
  • The first two are the fields: id and plot_id, which in this case are both numerical plot IDs
  • The last field is where the spatial information is stored and which is called geometry
  • Since this is point data each object is stored as a pair of x and y coordinates

  • This makes it a special kind of data frame that can be used by spatial tools

  • We can plot this data using a special geom, geom_sf
ggplot() +
  geom_sf(data = plots_harv)
  • We can also color vector data based on the values in the fields (or columns)
  • For example, our plots have two different types, “Tower” and “Distributed”
  • These are stored in the plot_type field
  • To color the points based on plot_type we add a mapping
ggplot() +
  geom_sf(data = plots_harv, mapping = aes(color = plot_type))
  • Just like in scatter plots this mapping tells ggplot to “color the points based on `plot_type”

Combining multiple spatial layers

  • We have the start This shows us the position of the plots, but it’s hard to learn much from this without some context
  • So let’s load another vector object that shows the boundary of the research site
boundary_harv <- st_read("data/HARV/harv_boundary.shp")
  • We can plot them together by adding two geom_sf layers in ggplot
ggplot() +
  geom_sf(data = boundary_harv) +
  geom_sf(data = plots_harv)
  • The order of layers is important because they will plot on top of one another
  • So if we’d plotted the plots first…
ggplot() +
  geom_sf(data = plots_harv) +
  geom_sf(data = boundary_harv)
  • We wouldn’t have been able to see them.
  • If we need to see through layers we can do this by setting the transparency using alpha
ggplot() +
  geom_sf(data = plots_harv) +
  geom_sf(data = boundary_harv, alpha = 0.5)