Chapter 3 Future Use

This chapter describes how to continue using the model in the future.

3.1 Technical Assets

3.1.1 GitHub Repository

The complete code for both model phases 1 and 2 is provided at the GitHub repository for this project (also linked in the top right corner). It is recommended that future changes to this model not made by Nelson\Nygaard be made in a forked version of this GitHub repository, so that changes can be tracked. The README.md on the GitHub page provides a description of each subfolder of the repository.

3.1.2 GitHub Pages Website

A relatively simple GitHub pages website was developed (using bookdown)to document this project (the website you are currently reading). It is part of the GitHub repository and be re-hosted at a forked version of this repository as needed.

3.1.3 Google Sheets Setup Spreadsheet

A Google Sheets spreadsheet was used to provide input values for the model (e.g., the trip type patterns) in a way that could be easily edited. That file is available here. To memorialize the settings for the model, the above link is a view-only link – edit permissions have been provided to relevant Translink staff via email.

3.1.4 PostgreSQL Database

The model code saves out results to a PostgreSQL database, and this database is pulled from when the results dashboard is assembled. This database is hosted on Amazon Web Services’s Relational Database Service. For now the data will be hosted on NN’s database in perpetuity, but a database dump or a machine image of the RDS server can be provided upon request for further modification.

3.1.5 Shinyapps.io Hosting

The results dashboard is a relatively simple Shiny application hosted at Nelson\Nygaard’s shinyapps.io server. It pulls from binary files assembled from the database in a separate script, identified in the README.md of the GitHub repository.

3.2 Scenario Modification/Addition

Scenarios can be added/modified by updating the Google Sheet were they were defined discussed above. The tabs that need to be modified are: - trip_type_pattern_defs: New trip type patterns can be added (or existing ones modified here.) Make sure that an appropriate identifier is created in the trip_type_patterns tab for new entries. - model_scenarios: Add a new scenario row (or edit an existing one) in this tab. The scenario ties together assumed ridership return percentage, social distancing, and the trip type pattern. If a new ridership return percentage is used, this should be added to the rr_scenarios tab as well so a new identifier can be created/used.

3.3 Addition of New Observed Data

Addition of new observed data should be done by modifying the data cleaning script (R/phase_2_load_clean/phase2_data_cleaning.R) to clean the new dataset and then merge it back into the overall observed dataset.

3.4 Model Changes

If there are any changes to the model code itself, a new version number should be created (currently using 0.3.0) and then a description of the changes in the new version should be added to README.md. The version number is used to separate model run results in the database – you wouldn’t want to mix simulations from different model versions.