GIS

" system designed to capture, store, manipulate, analyze, manage, and present all types of spatial or geographical data"

       -- wiki

 

Term coined in 1968

-Roger Tomlinson, Canadian regional planner

GIS

Relational database where index field is spatiotemporal reference

 

Generally, baselines are from topos, enhanced with satellite imagery / aerial data

Data representation

Vector

• continuous boundary; categorical value

 

Raster

• continuous value; systematic boundary

 

3D color images

Source: bgis.sanbi.org

Shapefiles

ESRI's file format

• Vector data (shape, poly, point)

 

• Set of files stored together in common directory

 

• Projection info stored separately from boundaries

 

 

 

Source: bgis.sanbi.org

Front-end

ArcGIS (ESRI)

Google Maps / Earth (KML)

Back-end

GDAL/OGR (Geographic data abstraction library)

• File import/export engine
• Raster-focused
• R access through rgdal: readOGR / writeOGR

GEOS (topological projections)

 

Shapefile architecture

• .shp = object boundaries
• .dbf = attribute data
• .prj = projection

Google maps

Data in kml files

KML = "keynote markup language"

Spatial vsn of XML

Map Projections

Systematic transformation of locations on a sphere to locations on a plane

 

Developable surfaces

• Surface that can be unfolded/unrolled without stretching, shriking, or tearing
• Cylinder, cone, plane

 

For sphere, some metric always gets distorted

Source: Geographic Information System Basics, Schmitz (2012)

Projection Strategy

1. Project to developable surface

 

2. Unroll surface into plane

Flattened to plane

Developable surfaces

Source: http://www.progonos.com

Projection considerations

What metric to distort?

• Area
• Shape
• Direction
• Bearing
• Distance
• Scale

 

Check out all these projections!

https://en.wikipedia.org/wiki/List_of_map_projections

https://www.washingtonpost.com/news/worldviews/wp/2015/08/18/this-interactive-map-shows-how-wrong-other-maps-are/

Coordinate Reference System

Examples

• Lat/Long
• Universal Transverse Mercator (UTM)

 

Elements

• Measurement framework (spherical coordinates or planimetric)
• Units
• Map Projection used

Package

rgdal ------------------------ brings GDAL/OGR functionality to R

 

spatial ---------------------- defines key spatial object classes

 

sp ---------------------------- transformation and drilling functionality

 

maptools ------------------ gets a lot of use, especially readShapePoly

 

rgeos ----------------------- interface engine connecting R with GEOS

                                   (C  API with lots of nice topological functionality)

Functionality

SPDFs

(Spatial Points (Polys, Etc) Data Frames)

• defined in spatial
• S4 objects
• SPDFObject@data gets you regular dataframe
• Heading slots to specify coordinate reference system

Convert dataframe into SPDF

1. Read in csv with coordinates stored in "Long" and "Lat"

MyData <- read.csv("<Path to data.csv>", header = T)

class(MyData) == data.frame

2. Use coordinates() function to specify coordinates

coordinates(MyData) = ~ Long + Lat

class(MyData) == SPDF

3. Specify coordinate reference system

latlong <- "+init=epsg:4326"

proj4string(MyData) <- CRS(latlong)

 

Read ESRI shapefiles

maptools

• readShapeSpatial("<Path to .shp>.shp")
  • Does NOT manage projection info
  • You have to do this by hand!!

 

rgdal

• readOGR(dsn = "<Path to layer's directory>", layer = "<Layer name no extension")
• imports projection information

Read KML files from google Earth

with maptools

GoogleEarthData <- getKMLcoordinates(kmlfile = "./Data/MyPoints.kml")

Points come in a list, so

GoogleEarthData <- do.call("rbind", GoogleEarthData)

Reprojecting

(Changing to another CRS)

Check current projection with

     proj4string(MySPDF)

 

spTransform() in sp to switch to a new projection

MySPDFReprojected <- spTransform(MySPDF, CRS(latlong))

(recall: we defined latlong earlier to be an object specifying a particular CRS)

Reprojecting

(to work with a google map tile)

Load RgoogleMaps

     require(RgoogleMaps)

 

Project data to decimal-degrees if it's not already there

MySPDFLatLong <- spTransform(MySPDF, CRS(latlong))

 

Recenter lat/longs relative to google map tile

MySPDFGoogle <- LatLon2XY.centered(MyMapTile,

                                                                             MySPDFLatLong$Latitude,

                                                                             MySPDFLatLong$Longitude,

                                                                             zoom = GoogleZoom)

 

Aggregating over layers

("Drilling down")

Goal:

Align two layers & extracts values from one layer that occur at spatial locations specified in the other

 

over() in sp

 

MyPoints <- readOGR(dsn = "./PathToPoints", layer = "MyPointsLayer")

MyPolys <- readOGR(dsn = "./PathToPolys", layer = "MyPolysLayer")

PolyValsAtPoints <- (MyPoints %over% MyPolys)$Value

Extracting overlap areas

Scenario:

Shapefile of polygons; extract overlap area for each pair

 

gIntersection() in rgeos

MyPolys <- readOGR(dsn = "./PathToPolys", layer = "MyPolysLayer")
# Extract geometries of the two focal polygons
Poly1 <- MyPolys$Polygons[[1]] # gets 1st polygon from MyPolys
Poly2 <- MyPolys$Polygons[[2]] # gets 2nd polygon from MyPolys

require(rgeos)
Poly1Poly2Overlap <- gIntersection(Poly1, Poly2, byid = F)

Building static maps

1. Import map tile from google maps

require(RgoogleMaps)

center.in <- c(45.358, -111.403)
MyMap <- GetMap(center = center.in,

                                 zoom = googlezoom,

                                 maptype = "terrain")

2. Project points and store in dataframe

MyPoints <- readOGR("./PathToPoints", layer = "MyPoints")

MyPointsGoogle <- LatLon2XY.centered(MyMap,

                                                                        MyPoints$Lat, MyPoints$Long, zoom = googlezoom)

MyPointsGoogleData <- data.frame(MyPointsGoogle)

3. Assemble plot

PlotOnStaticMap(MyMap)

points(MyPointsGoogleData$newX, MyPointsGoogleData$newY,

             pch = 16, col = "blue")

 

Building dynamic maps

RStudio's interface with leaflet

http://rstudio.github.io/leaflet/