GIS/GPS Chapter 7

Metadata

data about the data

What classifies metadata?

who collected, what was collected, where collected, when collected, why collect, how collected, scale of data, algorithms.transformations applied to data

Element of metadata

spatial data structure (raster/vector), projection, coordinate system, datum conversion/transformation, scale of original data, when created, how collected, database field names and properties, data quality/errors and any error reports, accuracy and precision of instruments used to collect data

FGDC metadata standards

identification information, data quality information, spatial data organization information, entity and attribute information, distribution information, metadata reference information, citation information, time-period information, contact information

Accuracy

extent that both attribute and positional data correspond to tehir real-world counterparts

Precision

exactness of measurements

Types of error in geospatial data

attribute error, positional error, topological error, topological geometric error, temporal error, interpretation error due to ecological fallacy, error due to modifiable areal unit problem

Attribute error

acceptable level of accuracy, 2 most common methods used to determine attribute accuracy, creation of an error matrix and the computation of overall accuracy, producer’s accuracy, user accuracy, and kappa coefficient of agreement, attribute RMSE, attribute logical consistency, attribute and spatial completeness

Acceptable level of accuracy

government agencies define their own minimum level of accuracy and always violating federal if not following it, some GIS users set their own ad hoc levels of accuracy

What are the 2 most common methods used to determine attribute accuracy?

random spot-checking, spatial sampling and only used when sub-populations are being used

Ground reference

what is being tested

Reading across row

what is has

Reading the column

what it should be

Attribute RMSE

do it for each row, tells you the amount of error, varies with the amount of data

What does a higher RMSE mean?

bad, lower is better

Attribute and spatial completeness

degree of which the data exhausts the universe of all possible items, same level of detail

Positional error

positional accuracy, maps accuracy standards with USES, american society for photogrammetry and remote sensing, FGDC

Positional accuracy

measure how close the geographic coordiantes of features in a spatial data later are to their real-world geographic coordiantes (horizontally and vertically)

USES

US national map accuracy standards (NMAS), drone and satellite with own system to work, stricter than FGDC standards

American society of photogrammetry and remote sensing

map accuracy standards for large-scale maps

FGDC

national standard for spatial data accuracy, social science data standards, measure error using RMSE

Topological errors

things associated with geometry

Temporal accuracy

how up-to-date is geospatial database, temporal relates to time, always dependent on short time-scales

Example of temporal accuracy

commuting home and trying to avoid heavy traffic, someone in the car has a navigation device with traffic information, updated frequently

Error visualization

visual error

Error propagation

half sand/clay, messed up with not making it all half and half, perpetuates to every layer and everything becomes wrong

Ecological fallacy

belief that all observations within an area will exhibit same/similar values for particular characteristic

Example of ecological fallacy

assume you live in school district where average standardized test score is 75/100, friend lives in school district where the average for the same test is 65/100, people assume that any given student in your school district would score higher on the test than from your friends school

Modifiable area unit problem

when smaller areal units are combined into fewer, larger units, the variation present in smaller units may decrease, causes variation in statistical results between different level of aggregation, degree of association/correlation between variables depends on size of areal units being analyzed, correlation between variables increases as the size of the areal unit increases

Example of modifiable area unit problem

COVID