flusight-evaluation report

This repository contains code to generate FluSight hospitalization forecast evaluations, and support for updating/publishing the reports via GitHub Pages. These reports are generated on Thursdays.

There are two ways to run the program: as a Docker container (the default, which is run on Amazon Web Services), or manually on a local machine.

Steps to run as a Docker container

Please see docker/README.md.

Steps to run manually on a local machine

• Clone/update your copy of https://github.com/cdcepi/FluSight-forecast-hub . Make a note of its location for the next step.
• Temporarily edit code/score_flusight_forecasts.R: 1/2) Uncomment the line that sets hub_path, setting it to the location of the repo in the previous step. E.g., hub_path <- "../FluSight-forecast-hub". 2/2) Comment out the lines after that line, starting with args <- commandArgs(trailingOnly = TRUE) through the end of the if statement.
• Generate scores: Run code/score_flusight_forecasts.R.
• Produce reports: Knit code/Evaluation_flu_hosp.Rmd. This will create a new .html file in the current directory.
• Rename and move the .html file to reports/YYYY-MM-DD_Evaluation_flu_hosp.html (where YYYY-MM-DD is current date). E.g., reports/2024-01-05_Evaluation_flu_hosp.html.
• Update reports/reports.json by running code/update-reports-json.py. This file is used by index.html to list all reports.
• Add, commit, and push the above two files under reports/ to this repo. This push will initiate rebuilding and publishing the website (it may take a few minutes).
• Check https://reichlab.io/flusight-eval/ to verify that it's been updated.