Skip to contents

FishSET GUI Quickstart Guide

Source: vignettes/FishSET_GUI_Quickstart_Guide.Rmd
FishSET_GUI_Quickstart_Guide.Rmd

FishSET GUI Quickstart Guide

FishSET is a set of statistical programming and data management tools developed to improve fishery modeling. The tools standardize data management and organization, enable the use of location choice models to provide input into fishery management, and provide various other modeling and visualization tools. The FishSET toolbox is provided as a set of R functions that can be run in an R console or here in this FishSET Graphical User Interface (FishSET GUI). This guide gives new users a practical path through the FishSET GUI.

Recommended path: Upload Data -> Select variables -> QAQC -> Explore the data -> Format Data -> Modeling -> Policy

This document includes basic instructions on how to use the FishSET GUI. Refer to the FishSET User Manual for information on installing R software and to the main page of this website for FishSET installation instructions.

What the GUI does

The FishSET GUI is a user-friendly interface that guides users through the steps necessary to run discrete choice models. The FishSET GUI requires no knowledge of coding or the R programming language as all steps to prepare and run discrete choice models can be run in a graphical user interface.

Before you start

  • Have your main and spatial data files ready. Port, gridded, and auxiliary files are optional.
  • Choose a FishSETFolder where project files can be saved.
  • For mapping or policy tabs, make sure your spatial data contain the required zone information.

What gets saved

  • Project data and edits are stored in a local SQLite database on the user's computer (data not shared with other users or FishSET developers) known as the FishSET database.
  • The database is housed in the project's directory within the FishSET R package directory (FishSETFolder).
  • Plots and tables are written to the project output folder.
  • Function calls and input values are written to the project logs.
  • The report template is stored in the project doc folder.
Open the GUI: run library(FishSET) and then run_fishset_gui() in R or RStudio. The main and spatial data tables are required; supporting files depend on your workflow.

Quick path through the GUI

Click any step below to expand the instructions.

Information across all tabs

All tabs include controls to close the GUI, manage database tables, record notes, and run an R expression. Within the GUI, the primary data frame is called values$dataset.

Upload Data

Load Files

  1. Click Change FishSET Folder to set where project files will be stored.
  2. Type a new project name or select an existing project and select it from the dropdown.
  3. Provide the primary data; port and auxiliary data are optional. If loading an existing project, select tables from the dropdown menus; otherwise, upload your files.
  4. Provide your spatial files. You can upload a standard spatial file (CSV, GeoJSON), or check the "Uploading shape file instead?" box to upload a multi-file shapefile. Gridded data is optional. (See Chapter 4 of the FishSET User Manual for more information on the different types of data.)
  5. Click Load data.

Tip: Confidentiality settings, log reset, table management, and full data refresh are also available in this tab. Refreshing restores the original loaded table unless another table is selected on the Upload Data tab.

Select Variables

Use the Main and Spatial data sections to assign IDs, coordinates, and dates, then click Save selected variables. Use the Create ID buttons if trip, haul, or zone IDs were not included in your uploaded dataset. These selections are saved to the database and only need to be selected when starting a new project.

QAQC

Data Quality Checks

Use the sidebar to check for missing values, empty variables, duplicate data, incorrect classes, outliers, and invalid spatial coordinates. Use Preview data and Summary table before applying corrections.

See the recommended data quality checks in Chapter 4 of the FishSET User Manual

Explore the Data

Switch between tables and temporal, spatial, and correlation plots to understand the loaded data. Generated plots and tables are saved to the project output folder.

For more details see Chapter 6 of the FishSET User Manual

Format Data

Compute New Variables

Use the radio buttons in the left sidebar to access various tools for creating new data columns. You can apply custom R expressions, calculate spatial metrics (like trip distances or centroids), aggregate your data (such as collapsing haul-level data into trip-level data), or generate group-level statistics (like quantiles, percentages, and cumulative sums). Follow the on-screen prompts for each selected tool to compute and save your new variables to the dataset.

Define Alternative Fishing Choices

Use this tab to define how the distance matrix between observed and alternative fishing locations should be calculated for your choice model.

  1. Set your Starting location (Port, Lagged haul, or Coordinates).
  2. Set your inclusion threshold (the minimum number of observations needed for a zone to be included as an alternative).
  3. Type a unique name for your matrix and click Create alternative location matrix to generate and save it to your project database.

Create Expected Catch Matrix

Use this tab to estimate the expected catch (or revenue) for the alternative fishing zones you defined in the previous step (zones where fishing could have happened but did not).

  1. Under Core Inputs, name your new matrix, select your catch variable, and link it to your saved Alternative Matrix (select a price variable as well if you want to calculate revenue).
  2. Adjust the Temporal Settings, Grouping & Calculation, and Data Handling sections to define how the moving averages should be calculated and how empty data should be treated.
  3. Click Create Expectations to calculate and save the matrix to your project database.

Format Model Data

Reshape project data into the long format required for modeling.

  1. Under Core Inputs, provide a unique name for this dataset, select which variables to retain, and select the Alternative Choice and Expectations matrices you created in the previous steps.
  2. Under External Data Integration, you can optionally join auxiliary or gridded datasets to your model data.
  3. Under Configuration, select how to handle missing data (imputation) and whether to calculate spatial distances between observations and zones.
  4. Click Run Format Data to process and save the formatted dataset to your project.

See Chapter 8 of the FishSET User Manual for more information on defining alternative choices, how expected catch/revenue is calculated, and model development.

Modeling

Model Design

Use this tab to construct the design matrices (the mathematical structures) required to actually run your discrete choice models.

  1. Under Source Data & Naming, provide a unique name for this design, select your newly formatted dataset from the dropdown, and choose your Model Type (Standard Logit or Expected Profit Model).
  2. Under Model Specification, type in your Utility Formula (e.g., chosen ~ catch + distance | vessel_length). You can also check the box to automatically scale your numeric covariates for better model stability. If you selected the Expected Profit Model, you will also need to provide a Catch Formula and select a Price Variable.
  3. Click Create Design Object to generate and save the matrices to your project.

Model Fit

Select a saved model design, provide an optional fit name, choose an EPM distribution when required, and click Fit Model. Use the model comparison controls to review saved fits side by side.

Model Cross Validation

Use this tab to perform k-fold cross-validation to assess the out-of-sample predictive performance of your model. Cross-validation can take a few minutes to run depending on your fold count. Once finished, scroll down to review the Overall Out-of-Sample Performance summary, as well as the detailed accuracy metrics and estimated coefficients broken down by each individual fold.

See Chapter 8 of the FishSET User Manual for more information on model development.

Policy

Zone Closure

Use this tab to create management scenarios where specific fishing zones are closed or have restricted catch limits.

  1. Select an Alternative Matrix from the dropdown to load your valid fishing zones onto the interactive map.
  2. Click directly on the map to select the specific zones you want to include in your closure scenario.
  3. In the Allowable TAC by Zone table, adjust the total allowable catch percentage (0-100%) for each zone you selected.
  4. Enter a unique Scenario Name and click Add closure to save the scenario to your project database.

Run Policy Simulations

Use this tab to simulate redistributed fishing effort and calculate welfare changes following implementation of the policy scenarios.

  1. Under the first column, select your baseline fitted Model and choose one or more Closure Scenarios to apply (or leave this blank to run a baseline simulation).
  2. Under the second column, set your Simulation Draws. If you are using a Standard Logit model, you must also select your Marginal Utility of Income (MUI) variable and check the box if it represents a cost. (This option is automatically hidden if you are using an Expected Profit Model).
  3. Click Run Simulation to calculate the expected welfare and save the results to your project database.

Summarize Effort Redistribution

Extract and visualize how fishing effort shifts geographically under policy simulations.

Summarize Welfare Impacts

Evaluate expected welfare changes or compensating variation resulting from policy simulations.

For more information on simulating policy scenarios see Chapter 9 of the FishSET User Manual.

Need more help?