Assignments‎ > ‎Labs (4930/6920/6915)‎ > ‎Old Labs‎ > ‎

Error Analysis with your Project

Important Dates

Relevant Discussion(s):

Introduction to Lab


All data and every analysis we do in GIS has some uncertainty associated with it. These can include unreliability uncertainties (e.g. measurement error, resolution, etc.), and structural uncertainties (e.g. methodological choices). Sometimes those uncertainties matter and sometimes they don't. That judgment call can only be made effectively when the purpose of the analysis is clearly articulated and the significance of uncertainty can be appropriately contextualized.

In some ways, this lab will expect more intellectual maturity from you than any of the previous labs. You will be asked to do an uncertainty assessment (which might include an error analysis) of some spatial analysis from your project. The instructions are deliberately somewhat vague and ambiguous (i.e. uncertain) as to give you lots of latitude in how you choose to approach this.

Ultimately, your job with this lab is to convince your audience that even in light of uncertainty in your input data and analysis methods, that your analysis is reasonable (i.e. fit for purpose), and how confident we can be in it. If you can not demonstrate this, at least you need to articulate why not and propose what you would need in order to improve the analysis in the future to the point it can be considered reliable and informative.

No specific guidelines are prescribed as to how you must convince your audience. You are expected to draw from concepts in this week's lecture on this topic and leverage existing tools to go through an uncertainty assessment and error analysis. 

Data for Lab

No data is provided for this lab as you are expected to use your project data.

Lab Objectives

The primary purpose of this lab is to force you to consider, quantify (to extent possible) and assess the significance of uncertainty in a spatial analysis you have performed for your project.

Meets Course Learning Outcomes 2, 3 & 4.

Task 1

  1. First consider (qualitatively) all the sources of uncertainty in the input data:
    1. Accuracy and Precision of input data set(s)
    2. Lineage of Dataset (i.e. metadata)
    3. Currency of Dataset
    4. Attribute Uncertainty
    5. Generalization
  2. Consider the methods you are using to perform spatial analyses (are there multiple ways to do the same thing?).
  3. Choose a specific spatial analysis from your project (there better be at least one you can use) and qualitatively assess the methodological uncertainties (i.e. choices). Could you do a sensitivity analysis to demonstrate the significance of this?.
  4. For the spatial analysis you have chosen to analyze, consider how the input data, methodology, and/or output can be quantitatively assessed (e.g.  a positional error analysis, error propagation, error estimate, etc.) and perform this error or sensitivity analysis.
  5. Report the results as stated below.
You are welcome to and expected to recycle any of the text, logic, and figures you produce in this lab in your final project.

What to Submit

Prepare a webpage(s) for this lab on your personal website for the course specifically for this lab, which:
  • Provides enough context for your project (e.g. refer to a project webpage on your site) so that the spatial analysis you are showing us makes sense,
  • Provides a write up and summary of the purpose, motivation, workflow, results, and interpretation of significance of the steps summarized above for task 1,
  • Shows some figures (maps, tables, calculations, diagrams, graphs, etc.), which illustrate the results of the error analysis and/or sensitivity analysis.
Remember:  Focus your submission on convincing your audience that your research (or at least one particular component of spatial analysis) is reasonable and meaningful by fully addressing the associated uncertainty in a quantitative and thoughtful way.

Make sure your lab conforms to the general lab submission guidelines. Submit a URL for this lab's webpage.

Additional Resources

See Lecture Topic Resources