Raster Data

Raster Data

Raster Data

Raster Data

Raster Data

Raster Data

• Altitude Data http://srtm.csi.cgiar.org/srtmdata/

• Landsat (many different bands) Data https://glovis.usgs.gov/app?fullscreen=0

• Nightlights Data http://sos.noaa.gov/datasets/Land/nightlights.html

Raster Data

• Satellites/planes capture a range of data depending on their sensors
  • Visible light, ultravoilet light, altitude, rain, gravity, etc.
  • Each ‘band’ is a separate raster image/layer
  • One band = greyscale
  • Multi-band = ex. Red, Green and Blue

Raster Data

• Not a shapefile, because any image file (TIFF, BMP, PNG, JPEG) can be used
  • Don’t think of colours, but numeric values for every pixel
• But we need to know where to map the pixels to the earth’s surface
  • So usually a ‘GeoTIFF’
  • Location defined by the corner pixels
  • Projection is still vital!
• Resolution determined by pixel size
  • Google Maps is about 10m - 50cm resolution, depending on location
  • Storage is a challenge

Raster Data

Visualizing Raster Data

• Raster data often makes no sense to the human eye
  • May appear all black, or all white
• We have to focus on the interesting range of the data to make a clear contrast
  • QGIS tries to do this for us
  • Or we can ‘manipulate’ it manually
• ‘Pseudocolour’ is also useful for single band images

Visualizing Raster Data

Visualizing Raster Data

Visualizing Raster Data

• Scale: -32768 to 32768

Visualizing Raster Data

Visualizing Raster Data

• Scale: -5 to 1731

Visualizing Raster Data

• Scale: -250 to 1380

Visualizing Raster Data

• Pseudocolour

Visualizing Raster Data

• From Altitude to Slope: Digitel Elevation Model

Slope Analysis

• Raster to Vector: Contours

Raster Calculations

• Which pixels (places) are above >1000m?
  • Use a raster calculator
  • Pixels which are ‘TRUE’ show up in white

Raster Calculations

• We can combine multiple rasters using basic maths
  • +, -, *, /
• Pixel values in each cell are combined to create a new raster

Raster Calculations

• Population change (as measured by night lights)

Raster Calculations

• Population change (as measured by night lights)

Raster Calculations

• Population change (as measured by night lights)

Rasters and Vectors

• To change the boundaries of the raster we can ‘clip’ it to a vector shapefile

Rasters and Vectors

• To change the boundaries of the raster we can ‘clip’ it to a vector shapefile

Rasters and Vectors

• Sometimes the information in a raster is too detailed
• We can simplify by averaging the raster values according to specific polygons
  • Average altitude in each province
  • Average income by neighbourhood
  • Average rainfall by country
• These are called ‘Zonal Statistics’
  • We can calculate lots of different statistics

Rasters and Vectors

• To change the boundaries of the raster we can ‘clip’ it to a vector shapefile

Rasters and Vectors

Machine Learning with Raster Data

Poverty Mapping

Google Earth Engine

Collect Earth