1 of 38

Earth Engine 101

Noel Gorelick

goo.gl/9sYsqW

2 of 38

Overview

3 of 38

What Is Earth Engine?

"Big Data" analysis and visualization platform

Inherently parallel system

Designed for scientists, not software engineers

Goals: make it easy, enable non-traditional users

4 of 38

What Is Earth Engine?

"Big Data" analysis and visualization platform

Inherently parallel system

Designed for scientists, not software engineers

Goals: make it easy, enable non-traditional users

Focused on society's biggest challenges

Deforestation Climate Change

Drought Conflict

Disaster Global Food Security

Disease Sustainability

5 of 38

9PB Public Data Catalog

Imagery

Landsat 4-8 7 bands, 30m

MODIS 250m Daily Global

Sentinel-1 10m SAR

Sentinel-2 12 bands, 10/20/60m

Geophysical

Digital Elevation

Land Cover

Surface Temperature, etc.

Weather Forecasts, Climate Models

+300 more analysis ready datasets

6 of 38

How Do I Use It?

Javascript API

Interactive Code Editor

Node.js*

Python API

Python module

Web Apps with Appengine

Jupyter Notebooks*

7 of 38

Concepts

8 of 38

Data Models

Feature

Line / Point / Polygon

List of Properties

TNC Ecoregions

9 of 38

Data Models

Feature

Image

Stack of Georeferenced bands

Each band has its own:

Mask, Projection, Resolution

A list of properties, including:

Date, Bounding-box

10 of 38

Data Models

Feature

Image

Collection

Bag of Elements

Table of Features

Directory of Images

Filter, Sort, Join, Map, Reduce

11 of 38

Map

Apply a function to each element of a collection

A "For Each" operation

Examples

Compute area of each feature
Cloud cover of each image
Mosaic for each month

12 of 38

Reduce

Apply a function to everything in a collection

"Aggregation"

Examples

Summed area over all features
Median-pixel composite
Train a classifier

13 of 38

Reducers in Earth Engine

8 ways to reduce

Image.reduce

Image.reduceNeighborhood

Image.reduceRegion

Image.reduceRegions

Image.reduceToVectors

ImageCollection.reduce

FeatureCollection.reduceColumns

FeatureCollection.ReduceToImage

40+ reducers

Reducer.allNonZero

Reducer.and

Reducer.anyNonZero

Reducer.count

Reducer.countEvery

Reducer.histogram

Reducer.intervalMean

Reducer.linearFit

Reducer.linearRegression

Reducer.max

Reducer.mean

Reducer.median

Reducer.min

Reducer.minMax

Reducer.mode

Reducer.or

Reducer.percentile

Reducer.product

Reducer.sampleStdDev

Reducer.sampleVariance

Reducer.stdDev

Reducer.sum

Reducer.toCollection

Reducer.toList

Reducer.variance

14 of 38

Reduce Bands

B1

B2

B3

B4

+ B5

103

15 of 38

Reduce Image Collection

Google Confidential and Proprietary

16 of 38

Reduce Neighborhood

B1

B2

B3

B4

B5

17 of 38

Reduce Region

B1

B2

B3

B4

B5

Dictionary

{

B1: 8.3,

B2: 14,

B3: 176,

B4: 1.6,

B5: 7

}

18 of 38

Reduce Regions

B1

B2

B3

B4

B5

	B1	B2	B3	B4	B5
F1
F2
F3

FeatureCollection

F1

F2

F3

19 of 38

Reduce To Vectors

labels

B1

B2

B3

B4

	B1	B2	B3	B4
F1
F2
F3
F4
F5
...

FeatureCollection

20 of 38

Band Math

+

=

21 of 38

Band Math

+

=

+

=

22 of 38

A Computation Example

Landsat

Image Collection

23 of 38

A Computation Example

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

24 of 38

A Computation Example

Surface reflectance

Image

Metadata

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

25 of 38

A Computation Example

Add NDVI band

NIR Band

Red Band

Surface reflectance

Image

Metadata

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

26 of 38

A Computation Example

Composite: max(NDVI)

Index Band:

NDVI

Add NDVI band

NIR Band

Red Band

Surface reflectance

Image

Metadata

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

27 of 38

A Computation Example

RGB Visualization

Bands: B4, B3, B2

min: 0.3

max: 0.8

Index band:

NDVI

Composite: max(NDVI)

Index Band:

NDVI

Add NDVI band

NIR Band

Red Band

Surface reflectance

Image

Metadata

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

28 of 38

A Computation Example

Map: Reproject to Mercator

EPSG:3785

Scale: 30m

RGB Visualization

Bands: B4, B3, B2

min: 0.3

max: 0.8

Index band:

NDVI

Composite: max(NDVI)

Index Band:

NDVI

Add NDVI band

NIR Band

Red Band

Surface reflectance

Image

Metadata

Filter by date

2001-01-01

2003-12-31

Landsat

Image Collection

29 of 38

Tiling