Video #20 (Uncertainty) Flashcards
Why are errors important?
They can invalidate our analysis
Where is the largest source of errors typically?
In the data aqusision faze
What is accuracy?
How accepted the values of data are represented
One of the downsides of highly accurate data is…
It is expensive to produce or record
Because of the inevitable errors and inaccuracies, we standardize accuracy using…
A margin of error
Precision is…
How specific and how deep the data is (0.200 is more precise than 0.2)
Precision in GIS usually entails…
Greater detail of a map
True of false: Extremely precise data is more accurate data
False; Precision =/= Accuracy
What is the level of precision?
The accepted precision value which is dependant on what you are surveying (width of road can be less precise than width of Canada)
What is data qualtiy?
The accuracy and precision of the data collected.
Data quality is assessed in…
Data quality reports
What are the types of errors?
PAC
1) Positional error
2) Attribute error
3) Conceptual error
Positional error is…
The error of position in a GIS, possibly by using the wrong scale or coordinate system
Attribute error is…
The non-spatial analysis of data may be inaccurate (single family home may actually be a condo) or imprecise (home instead of condo or lacking what floor a person may live on in a condo)
Conceptually error is…
The use of inappropriate categories or data used (having elevation as a dataset while analyzing income in proximity to city centres)
What is the issue with the “How Many Lakes in Finland?” study?
The data was extremely precise which raised questions about if it was valid or accurate, and the definition of a lake wasn’t given (is a pond a lake?)
Misclassification is…
Using information to represent a phenomena in an ineffective way. The % of voters would be an ineffective way of classifying or categorizing GDP
How is the threshold of data quality determined?
By the user’s instinct. It should be appropriate for the project
The GIS software can provide some sources of error itself. An example of this is…
Fuzzy borders or impreciseness (think walking distance on my final project)
What is the recognition of error?
The levels of error that are present in the GIS, which should be acknowledged by the author
What are the 3 major sources of error?
1) Obvious errors
2) Natural variation errors
3) Processing errors
Why are obvious and natural variation errors more easily detectable than processing errors?
Processing errors are subtle and hard to identify, beginner users may be unfamiliar with errors that can occur in processing
What are some examples of obvious errors?
Age of data, map scale, relevance of data, accessibility etc.
What are the 2 issues with the obvious-age of data error?
The space it is trying to represent has changed. A map of Vancouver in the 1930s will be inaccurate today. Standards have also changed since the map was produced
What is the obvious-areal cover error?
Some data may be left out of skewed
What is a obvious-map scale error?
The scale used varies how fine your data may be. A scale of 10:1 will be more accurate than 400,000:1 in showing a house layout
What is a obvious-sample size error?
A greater sample gives more precise results so a small sample isn’t as appreciated
What are some processing errors?
Numerical errors, topological analysis, classification and generalization, digitizing
What are processing-numerical errors?
Errors that may occur because some hardware or software may not be well equipped to deal with what the user is inputting or rounding in processes
What are processing-topological errors?
Errors where the process doesn’t recognize certain topography or recognize what the user is inputting
What are processing-classification errors?
Errors where the classification of data is unclear
What are processing-digitizing errors?
Errors where the software recognizes coffee spills on physical maps as polygons
What is data consistency and why is it important?
Data consistency is the consistency of measurements and scale in data, and it’s important because it sets a standard of measurement and data can blend well together
What is the problem with missing data?
Missing data can change the narrative of the data by misrepresenting a phenomena
What could a reason some road data is missing?
The definition of a road could change between sources
What is meta-data?
Data about data. Data includes lineage which is the meta-data
What are the six important things included in meta-data?
1) Specification of sampling methodologies
2) Definition of terms
3) Measurement specification
4) Documentation of classification system
5) Data model methodology and history
6) Purpose of study
What should you do in reaction of error and uncertainty?
Learn why it is inaccurate, express it and engage with it
