North Atlantic Whale Project

• Summer project from 2020. Worked with Professor Steve Wang and Horace Shew, all from Swarthmore College. Previous contributors are mentioned in the Rcode.
• In this project, we predict the population size of North Atlantic Whales, an endangered large species.
• Data is pulled from NARWC

Parsing Code

• First half of the project, where we transform the dataset model from sighting-level to whale-level
  • Sighting-level: each entry is a sighting of a whale, thus there could be multiple entries for the same individual whale
  • Whale-level: each entry is unique to a whale, identified by callouses and patterns on their back

Rcode

• Main portion of the project. Converts sighting-level dataset to a whale-level datset (i.e. so that each entry is unique to one whale),
• Logistic regression model that would adjust individual whales' pre-calculated survival rates.

Gap time code v1

• Gap time code v1 analyses the average gap time of all whales (thus, for whales that are seen again)

Gap time code v2

• Gap time code v2 is different than v1 in that:
  1. It counts the number of individual whales observed in the year t
  2. Then, it looks at t+1 through t+5 years and checks if the whales from year t are observed again.

These two methods are efforts to more appropriately calculate probaliveold, a vector of survival rates of individual whales (identified by unique whale IDs) that are later to be adjusted by our model.