MaptimeNYC 1.24.2017

New Year's Resolutions:

Working w/Satellite Data

January 24th, 2017

Follow along here: https://github.com/MaptimeNYC/working-with-satellite-data

Geographic Data 101

2 Types:

Vector
Raster

Vector

Vector represents real-world objects (buildings, rivers, streets, GPS points). Geometries are defined by the vertices that make points, lines, & polygons.

Common Vector Data Formats

GeoJSON
Shapefile

Vector Example

Raster

Rows and columns of cells (or pixels) create a grid of data
Cells hold values (temperature, elevation, red/green/blue color)
Grids organized into bands (channels) that can overlay multiple different types of values
Size of each cell makes the dataset's resolution

Common Raster Data Formats

GeoTiff
...?

Digital Elevation Model

Temperature Data

MPG Ranch Restoration Map

Landsat

Joint effort by NASA and USGS
Conceived in 1965
Landsat 1 Launched 1972
Public Earth science datasets for 44+ years!

Landsat Mission

“Landsat 9 [will] propel the program past 50 years of collecting global land cover data,” said Jeffrey Masek... “That’s the hallmark of Landsat: the longer the satellites view the Earth, the more phenomena you can observe and understand.”

NASA - A Landsat Timeline

Landsat8

Launched February 11, 2013
11 bands
Each band from a different part of the electromagnetic spectrum
Visible spectrum < half of Landsat's data
Surveys entire Earth every ~16 days
Most bands are 30m resolution, Thermal-Infrared (bands 10,11) 100m resolution

Google Earth Engine

"Platform for planetary-scale environmental data analysis."
"May only be used for development, research, or educational purposes."
Includes a codepen-like script editor (Python or JS) for easily accessing Earth satellite datasets and the processing power of Google's cloud resources

Getting Started with Earth Engine

Sign Up for Earth Engine Developer Account (takes a few days)
Once you have an account, all the code that follows is meant to be entered in the Earth Engine Code Editor

Adding Data

Use this Google data catalog to browse available datasets

In EE code editor, add the dataset to the map document:


							/* 1.  Add satellite data to map  */
							Map.addLayer(
							  //  add collection of raw Landsat8 data
							  ee.ImageCollection('LANDSAT/LC8_L1T')
							);

Zoom to Location

(below previous code in EE code editor)


							/* 2.  Center map on Indian Point (longitude, latitude, zoom) */
							Map.setCenter(-73.954661, 41.269739, 15);

Filter Datasets


							/*  Filter by a vector(!!!) geometry */
							var indianPoint = ee.Geometry.Point(-73.954661, 41.269739);

							var filteredCollection = ee.ImageCollection('LANDSAT/LC8_L1T')
							  .filterBounds(indianPoint);

							Map.addLayer(filteredCollection);

Filter Datasets By Time


							/*  Filter by a vector(!!!) geometry */
							var indianPoint = ee.Geometry.Point(-73.954661, 41.269739);

							var filteredCollection = ee.ImageCollection('LANDSAT/LC8_L1T')
								.filterBounds(indianPoint)
								.filterDate('1990-01-01', '1990-05-31');

							Map.addLayer(filteredCollection);

Visualizing Vegetation

By combining different bands of Landsat data, you can visualize the otherwise non-visible data

See here for examples of band combinations


							/*  Add band visualizing 5-4-3 (infrared) */
							Map.addLayer(
								ee.ImageCollection('LANDSAT/LC8_L1T'),
								{'min': 6000, 'max': 18000, 'bands': ['B5', 'B4', 'B3']});