NHS COI

Presenting structured and raw data we received in response to requesting Conflict of Interest registers from every Trust in England.

Implementation notes

The script generate_data.py does the following:

• NHS Trusts. This file propercased, headers added, turned into trusts.csv
• Google Sheet originally created by Harriet. It contains 3 worksheets; we're interested in two:
  • The "Responses" worksheet contains questions and answers for individual Trusts (e.g. did they respond on time, etc). We convert this to a CSV hospitality-coi-raw.csv. The data from this CSV is displayed on the page for an individual trust. It also contains a column with a reference to a DOI from Figshare, with which we can refer to the document(s) returned by Trusts in response to the FOI requests.
  • The "Transparency Score" worksheet is a hand-normalised version of a subset of the same data, with a Total Score column. We add a ranking column, and convert this to hospitality-coi-scores.csv. This data is displayed on the home page.
• hospitality-coi-questions.csv - manually curated, single-row spreadsheet with friendly names for the titles at the top of the first worksheet of the Google Sheet

The two trust worksheets are joined by the "org code" column which must be consistent between them and the trusts.csv file. The questions worksheet is joined by the respective column headings.

Jekyll generates static pages by running the plugin code at _plugins/trusts.rb, which iterates over the contents of the three CSVs described above.

Docker commands

# Build the site
docker compose up build

# Remove the built site (for a clean rebuild)
docker compose up clean

# Run a local development server
docker compose up dev