Tests#

About a week after each lab, there will be a short (30-minute) written test on the lab topics.

The main purpose of these tests is to provide an incentive to work through the labs and readings yourself, without relying too heavily on AI. The tests - and the grading - are designed to be straightforward for students who have completed the labs and readings and understand the material.

Tests will be scheduled during the first half of class on the day indicated in the schedule.

I use Gradescope and will provide printed sheets with the questions and space to write your answers. All you need to bring is a pen and your brain.

Most questions will call for short written answers that test basic understanding. I might ask you to identify an error in a short code snippet or write pseudocode illustrating how a particular step works. I will not ask you to write flawless, fully functioning code.

Lab 1 test#

6%, based on your final Lab 1 performance

There is no proctored test for Lab 1. Instead, you can fill out a self-assessment at home (Lab 1 practice test) to get a sense of how you perform and what future tests will look like (type and count of questions). Your final Lab 1 performance replaces the test result.

The practice test covers your understanding of the Python geospatial stack introduced in Lab 1 and of the main introductory readings (Jenkins vs. Brown, Goulder & Kennedy). It has questions on:

  • pandas indexing/selecting and geopandas operations.

  • Coordinate reference systems: which ones to pick for different tasks.

  • Operations such as zonal statistics and raster reprojections.

  • Aligning & interpreting the maps you made.

  • Key components of conservation planning (esp. Brown’s perspective).

  • Basics of valuing ecosystem services (Goulder & Kennedy).

Lab 2 test#

7%

Tests your understanding of conservation planning with Marxan, including problem framing, data structure, optimization process, and interpretation of responses to changes in inputs. Mostly short-answer questions with small coding exercises.

Lab 3 test#

7%

Tests your understanding of explanatory and predictive modeling with statsmodels and scikit-learn, cross-validation, the structure of tree ensemble models, and policy targeting simulations, Newburn-style.

Lab 4 test#

5%

Tests your understanding of rigorous impact evaluation, counterfactual thinking, selection bias, confounders, and Mahalanobis covariate matching.