brahmwg
commented
4 months ago This potential deliverable was suggested by another of our researchers. I will communicate with them to get as much detail as possible and respond back here with the info. I think the idea is to have a map using data from microtrolling to display the stock composition of fish caught in different areas of the strait of georgia. Partners in DFO (department of fisheries and oceans) want to be able to interact with a map to select different shapes or regions and see the stock composition of that selected area. In the interim, focus on the other deliverables that you can make progress on, while I await more info from our researcher.
Updated description from researcher:
Salmon managers need flexible tools to estimate how different rivers populations are distributed in ocean regions where they are fished. Particularly, there is interest to incorporate uncertainty in fishery regulation decisions. Outcome: A R Shiny app with an interactive map (e.g., with the plotly package), with points representing salmon catch locations and associated genetic population assignment (which includes additional uncertainty). The user can define a "gate" or "lasso" (e.g., with the tidygate package or inside plotly), and output a dataframe of all data points with a variable defining which side of the gate each salmon data point is. With this data frame, work with UVic/PSF researchers to automate modelling of the proportion of each population found inside or outside of the gate (multivariate tweedie distribution; Thorson et al. 2022 Ecology).
Data: I can provide an example file (salmon ID, longitude, latitude, stock assignment probabilities). Using Adult Salmon Diet Program data, so that this is an example application rather than a final product.
https://github.com/brahmwg/Bottlenecks_MDS_Capstone/blob/main/deliverables/interactive_stock_composition_estimator
Not sure what this is XD, we'll ask in the meeting