The "ANALYSE" part of the toolbox is where you can modify or extract information from the data you selected in "INPUT".

These are the options available within the ANALYSE part of the toolbox.  They are separated into Calculate and Classify. 

Calculate:

Classify:

Instructions for using the analytics

CALCULATE blocks are added to container blocks, usually the dataset block. CLASSIFY blocks are container blocks and other blocks are inserted into these. 

The following sections describe how to use each of the available blocks. 

TABLE OF CONTENTS

• CALCULATE
  • Add Index
  • Calculator
• CLASSIFY
  • Create Supervised Classification
  • Create Unsupervised Classification

CALCULATE

Add Index

• This block provides a list of pre-defined indices for use on your data.
• This is a shortcut to calculating the indices in the Calculator block.
• The options will dynamically change depending on which dataset has been selected. 
• When it is placed within a workflow for which there are standard indices, the block will give you the choice of a category, and then an index. 
• For optical data, the categories, and relevant indices are:
  • Category: Burn
    • Index: BAIM. This is a Burned Area Index adapted for MODIS bands and uses NIR and short-wave IR (SWIR).
    • Index: NBR. The Normalised Burn Ratio is a normalised difference of NIR and short-wave IR (SWIR).
    • Index: NBR2. Normalized Burn Ratio 2 modifies the NBR to highlight water sensitivity in vegetation and may be useful in post-fire recovery studies. 
  • Category: Urban
    • Index: IBI.  Index-Based Built-Up Index is a combination of three thematic indices: the soil adjusted vegetation index (SAVI), the modified normalized difference water index (MNDWI) and the normalized difference built‐up index (NDBI).
    • Index: MBI. Modified Bare Soil Index uses short-wave infrared (SWIR) and near-infrared (NIR) bands.
    • Index: PISI. Perpendicular Impervious Surface Index. This uses blue and NIR bands, and is thought to give better contrast with vegetation and bare soil.
  • Category: Vegetation
    • Index: EVI. Enhanced Vegetation Index.   This includes the blue channel to improved upon the NDVI.  (The values used for the constants are the most commonly used: g=2.5, L=1.0, C1=6.0, C2=7.5.) 
    • Index: NDVI. Normalised Difference Vegetation Index. This is the classic index that is the standard default for any vegetation mapping.  
    • Index: SAVI. Soil Adjusted Vegetation Index. This includes a factor to account for background soil contributions (and so is helpful for agricultural studies).  (L=1.0).
    • Index: SR. This is a simple ratio of NIR/R. 
  • Category: Snow
    • Index: NDSI.  The Normalised Difference Snow Index. It picks out high visible reflectance (from the green band) with low short-wave IR (SWIR) reflectance, which is characteristic of snow (rather than clouds, which are high in both). 
    • Index: NDSII. Normalised Difference Snow Ice Index. This uses a visible (green) and NIR band to create a normalised difference index that picks out both snow and ice.  
    • Index: NDSInw. Normalized Difference Snow Index with no Water.  This uses NIR and SWIR bands to ensure that open water is not accidentally included in the snow index. The constant beta=0.05.
    • Index: SWI. The Snow Water Index is a combination of the green, NIR and SWIR bands and is good for reducing the impact of neighbourhood classes like cloud, soil, vegetation, and water. It is particularly good in mountainous areas. 
• For radar data, there is only one category, radar, which offers the following indices:
  • Index: DPDD. Dual-Pol Diagonal Distance. (VV + VH)/2.0^0.5
  • Index: RFDI. Radar Forest Degradation Index. (HH - HV)/(HH + HV)
  • Index: VV VH D. The VV-VH Difference.  
  • Index: VV VH R. The VV-VH Ratio. 

IMPORTANT NOTE ON INDICES:  You should think carefully about how the best way to use your indices, and whether it comes before or after the Create Composite block.

Since creating a composite aggregates data using some kind of an average it is sometimes better to generate the index for each image, and then average across each pixel.

Other times, you might want to average the pixels in the composite first, and then use the index.

Typically, if you expect the target to remain constant, then doing the index afterwards is OK.

Calculator

• The Calculator block allows you to apply arithmetic operations to available bands within a workflow.  This is sometimes also referred to as Band Math.
• When you click on the options icon you will get a pop-up window that allows you to construct the expression you want to implement.
• Saved bands from other workflows can also be used in the arithmetic expression.  
• This block also allows thresholding so that you can select portions of a dataset that meet certain criteria.
• For a full description of how to use this block see the article How To Apply Your Own Algorithms.

CLASSIFY

Create Supervised Classification

• This block performs a cluster analysis using machine learning using the training data that you supply.  This means that it looks at the pixels associated with each class in the training data and tries to identify data properties that they have in common.  
• The number of classes is defined by the training data.
• Supervised classification is used when we have training data (e.g. field data) that can be used to assign each pixel to a specific (unique) class. The resulting classes represent pixels that have similar data characteristics that are labelled according to the training data.
  
• Training and/or validation data can come from a variety of sources.  This block allows you to select areas on the map as training data.
  
• For a detailed example of how to use this block see the article How To guide on classification .
• The classification methods available are:
  • CART (Classification and Regression Trees) is a form of decision tree model. 
  • Random Forest is another decision tree method that builds multiple decision trees and then decides which is best. See Breiman (2001) 
  • Minimum Distance.  This is the simplest method -- it assigns each pixel to the nearest class.
  • SVM (Support Vector Machine) is a form of supervised learning model. See Burges (1998).
• Each of these methods has one or more parameters that can refine the algorithm in some way.  Most Earth Blox users will be happy to use the defaults for most values, but you can experiment by selecting the options icon. Without a priori information about the physical nature of the problem, optimal parameters are difficult to identify in advance. 
• The numeric results of the classification are automatically tabulated in the DASHBOARD tab. 
• An accuracy assessment is also provided automatically. 

Create Unsupervised Classification

• This block performs a cluster analysis using machine learning.  This means that it looks across all the available input data and looks for where pixels seem to cluster together with similar properties.
• Each of these clusters is assigned a class. 
• Unsupervised classification is used when we do not have training data (e.g. no field data) that can be used to assign each cluster to a specific (unique) class. The resulting classes are therefore merely representative of pixels that have similar data characteristics.
• For an example of how to use this block see the article How To guide on classification [ARTICLE STILL UNDER CONSTRUCTIO.
• Earth Engine uses the Weka software to implement unsupervised classification using optimised machine learning. Weka refers to a collection of algorithms made available by Waikato University. It optimises the machine learning code.
• The classification methods available are:
  • Cascade K Means.
  • Cobweb.
  • K Means.  
  • LVQ
  • X Means
• Each of these methods has one or more parameters that can refine the algorithm in some way.  Most Earth Blox users will be happy to use the defaults for most values, but for the "number of classes" variable it is often worth experimenting. Methods such as LVQ and K Means will ask for an exact number of classes, which is appropriate if you already have some idea of how many classes you expect to see in the area of interest. Other methods, such as the Cascade K Means and X Means, give you the option to specify a minimum and a maximum, giving the algorithm some flexibility in the analysis of the clusters. 

We always track the feedback you give us so that we can improve how we help you. Please let us know if you found this one helpful.