pontiPy 2.6

Automation of Pontius Matrix

Python Library developed and created to automate the generation and analysis of the PonitusMatrix.
Version History:

• Version 1.0 (Multiple Categories)
• Version 1.1 (Multiple Categories, Quantity, Exchange, Shift, Size, Difference)
  Python Version: 3.7
  Latest Publish Date: November, 2020

Dependencies & Usage

• Download: pip install -i https://test.pypi.org/simple/ pontiPy==2.6
• Required libraries for pontiPy usage

import pandas as pd
from pontiPy import pontiPy

• Creating a Dataframe from an inputted Sample & Loading the pontiPy package

initialDataFrame = pd.read_csv('sample.csv', index_col= 0)
display(initialDataFrame)
NewDataFrame = pontiPy(initialDataFrame)

Functions

Available functions in pontiPy Library

Matrix

• Displays inputted matrix
• Generates PontiusMatrix Contingency Table
  This will generate the final output contingency table

Arguments

• Generates Initial Contingency Table = matrix()

Examples

• Displaying the sample data through the contingency table function

display(NewDataFrame.matrix())

Size

• Function to Compute Size for all or one Category k
• Axis must be specified when category k is specified
• Determines if row or column sum for category k will be returned

Arguments

• No category specified = size()
  Size of Extent
• Category k specified = size(k)
  Size of category k
  a) Axis 'X' = Size of category k in X (row sum): size(k, X)
  b) Axis 'Y' = Size of category k in Y (col sum): size(k, Y)
  c) No Axis specified = Size of category k
• Category k Specified with TRUE = size(k, X or Y, TRUE)
  Dictionary listing the size of other categories contained in category k (includes hits)
• Category k Specified, Total = size(k, Total = TRUE)
  Returns the Total. Only works if a category is provided. If True, returns the total size (x & y) for category k

Example

print('Size of Extent:', NewDataFrame.size(), 'Hectares') 

Size of Extent: 25662 Hectares

print('Size of Category 1 in X:', NewDataFrame.size(0,'X'),'Hectares')  

Size of Category 1 in X: 2296 Hectares

print('Size of Category 1 in Y:', NewDataFrame.size(0,'Y'),'Hectares')

Size of Category 1 in Y: 2144 Hectares

Difference

• Function to compute difference for all or one category

Arguments

• No category specified = difference()
  Equation = (Total Size - Hits)
• Category k specified = difference(k)
  Equation = (Size-(2Hits)) For That Category

Example

print('Total Difference:', NewDataFrame.difference(), 'Hectares')  

print('Difference for Category 1:', NewDataFrame.difference(0), 'Hectares')  

Hit, Miss, and False Alarm

• Functions to compute Hits, Misses, and False Alarms

Arguments

• No Category specified = hit(), miss(), false_alarm()
  Sum of Total Hits, Misses, or False Alarms
• Category k specified = hit(k), miss(k), false_alarm(k)
  Hits, Misses, or False Alarms for Category k

Example

print('Total Hits:', NewDataFrame.hit(), 'Hectares')  

Total Hits: 3553 Hectares

print('Total Misses:', NewDataFrame.miss(), 'Hectares')  

Total Misses: 8735 Hectares

print('The Total False Alarms:', NewDataFrame.false_alarm(), 'Hectares')  

The Total False Alarms: 8735 Hectares

Quantity

Function to compute quantity between all or one category
Returns as a dictionary

Arguments

• No Category Specified = quantity()
  Total Quantity Disagreement
  Equation = [(Sum of quantity(k) for all k) / 2 ]
• Category k Specified: = quantity(k)
  Quantity Disagreement for k: |M-F|
• Category k Specified with TRUE parameter): = quantity(k,TRUE)
  Quantity is returned with Miss or False Alarm labels:
  If M-F is positive = Miss Quantity
  If M-F is negative = False Alarm Quantity
  If value = 0, then Quantity is 0

Example

print('Total Quantity Disagreement:', NewDataFrame.quantity(), 'Hectares')

Total Quantity Disagreement: 76 Hectares

print('Quantity Disagreement for Category 1 w/ label:', NewDataFrame.quantity(0, True), 'Hectares')

Quantity Disagreement for Category 1 w/ label: {'False Alarm': 152} Hectares

Exchange

Function to compute Exchange between ALL, ONE or TWO categories

Arguments

• No Category Specified = exchange()
  Sum of total exchange is returned
  Total must be false
• If total is False and 1 category is specified: = exchange(TRUE,k)
  Return is exchange for that category with all other categories + a total value in dictionary
• If Total is True and 1 category is specified: = exchange(FALSE,k)
  Return is total exchange for that category
• If 2 categories are specified (Total must be false): = exchange(FALSE,k,k)
  Return exchange between 2 categories

Example

print('Total Exchange Disagreement:', NewDataFrame.exchange(), 'Hectares')

Total Exchange Disagreement: 0 Hectares

print('Exchange between Category 1 and 2:', NewDataFrame.exchange(0,1), 'Hectares')

Exchange between Category 1 and 2: 0 Hectares

Shift

Function to compute shift between ALL or ONE categories

Arguments

• No Category Specified = shift()
  Total Shift Disagreement
  Equation = [(sum of shift(k) for all k) / 2 ]
• Category k specified = shift(k)
  Shift Disagreement for k
  Equation = (difference(k) – quantity(k) – exchange(k))

Example

print('Total Shift Disagreement:', NewDataFrame.shift(), 'Hectares')

print('Shift Disagreement for Category 1:', NewDataFrame.shift(0), 'Hectares')

Further Information & Contact

• Library Information:
  Priyanka Verma, [e-mail]

• Metric Methodolgy:
  Robert (Gil) Pontius, [e-mail]
  See Metrics that Make a Difference, Chapter 4
  Link to Chapter 4

Acknowledgements

Dr. Robert Gilmore Pontius Jr created the first version of this workbook in 2001. Pontius has revised this workbook several times, and each revision has a larger number for the suffix of the filename. Pontius created version 42 in 2019.

Visit www.clarku.edu/~rpontius for publications on this workbook's methods. Specificaly, see: Pontius Jr, Robert Gilmore and Ali Santacruz. 2014. Quantity, Exchange and Shift Components of Differences in a Square Contingency Table. International Journal of Remote Sensing 35(21): 7543-7554.

Pontius Jr, Robert Gilmore. 2019. Component intensities to relate difference by category with difference overall. International Journal of Applied Earth Observations and Geoinformation 77: 94-99.

Pontius Jr, Robert Gilmore and Marco Millones. 2011. Death to Kappa: birth of quantity disagreement and allocation disagreement for accuracy assessment. International Journal of Remote Sensing 32(15): 4407-4429.

Aldwaik, Safaa Zakaria, Jeffrey A Onsted, and Robert Gilmore Pontius Jr. 2015. Behavior-based aggregation of land categories for temporal change analysis. International Journal of Applied Earth Observation and Geoinformation 35(Part B): 229-238.