Skip to main content
Investigative Journalism Foundation

Appointments Database Handbook


About Orders in Council

Throughout this handbook, the definitions and explanations use language appropriate to the federal level; however, the same explanations apply at the provincial level, with the premiers substituting for the prime minister and the lieutenant-governor substituting for the Governor General.

Orders-in-council (OiCs) are a type of statutory instrument issued nominally by the Governor General acting for the monarch and on the advice of the Privy Council. However, in practice, OiCs are drafted and passed by the cabinet.

OiCs are issued whenever the prime minister and cabinet make a governor-in-council appointment. In short, governor-in-council appointments are formal appointments made in the name of the Crown on the advice of the government. 

Sometimes, these types of appointments cover positions where the individual will carry a fair degree of decision-making power and will have the authority to act independently from the government. 

For example, some of the positions that fall under the umbrella of governor-in-council appointments include judges, ambassadors and high commissioners, members of tribunals and boards, ombudspersons, commissioners and executives of Crown corporations. 

Other less consequential positions might be specified as requiring appointment by the governor-in-council under federal or provincial legislation, such as the prime minister’s housekeeper and chef under the Official Residences Act.

Sometimes, governments have issued OiCs to announce other types of appointments and, therefore, our database may also feature a limited number of non-governor-in-council appointments.

About our database

Orders-in-council are published online by the federal government and most provincial governments. The IJF collects all OiCs announcing appointments from each of these jurisdictions and then extracts the relevant information in a structured format.

Our appointments data include the federal, B.C., Alberta, Saskatchewan, Manitoba, Ontario, New Brunswick, Nova Scotia, Newfoundland and Labrador and Prince Edward Island governments. Our earliest records date back to 1990, and we add more every day. 

Our consolidated appointments database has two advantages for its users:

  1. Unified formats: Compare appointments reported disparately across Canada.
  2. Cleaner data: Follow patterns made possible through newly structured data collected from previously unstructured and/or unsearchable datasets.

Data dictionary

The detail page, when clicking through an entry on our appointments database, shows all non-empty data fields of a record. A user can also download the CSV of the results of a search to view every column.

Name: Appointee name stated in the OiC.

Position: Position of the stated appointee.

Organization: Entity that employs the appointee.

Location: Where the appointee will be located for the duration of the appointment period.

Region: The source region of the data, one of the regions outlined in the Catalogue of Sources.

Start/End/Posted Date: Respectively, the start and end date of an appointee’s appointment, and the date the appointment was confirmed through an OiC.

Acts: Associated law, passed by Parliament, that grants the authority to make the respective appointment.

Remuneration: Salary/compensation to the appointee for the appointed position or role. 

Reappointed: An appointee can be explicitly reappointed to a role or position they have previously occupied.

OiC ID: Unique government label for each OiC, used to identify and distinguish every order-in-council.

Catalogue of Sources

The appointments database is composed of 10 datasets covering the federal government and nine provinces. This table shows the date range for each jurisdiction provided in our database.

Jurisdiction

Source

Date range

Update Frequency

Federal

Orders in Council online database (Privy Council Office)

1990–01-03 - present

Daily

British Columbia

BC Laws online database (King’s Printer of British Columbia)

1990-01-01 - present

Every 100 orders

Alberta

Orders in Council online database (King’s Printer of Alberta)

1998-06-03 - present

Daily

Saskatchewan

Orders in Council online database (Sask. Cabinet Secretariat)

2007-11-21 - present

Daily

Manitoba

Orders in Council webpage (Man. Executive Council)

2007-01-03 - present

Monthly

Ontario

Orders in Council online database (King’s Printer for Ontario)

2016-07-06 - present

Daily

New Brunswick

Orders in Council online database (N.B. Executive Council Office)

2009-01-04 - present

Daily

Nova Scotia

Orders in Council online database (N.S. Executive Council)

2010-01-01 - current

Daily

Newfoundland and Labrador

Orders in Council online database (Cabinet Secretariat)

2003-01-03 - present

Daily

Prince Edward Island

Orders in Council online database (P.E.I. Executive Council Office)

2003-01-07 - present

Daily

Four jurisdictions were excluded as their orders-in-council are only published through their respective provincial gazettes and not via an online database, making data collection prohibitively difficult. These jurisdictions are Quebec, Yukon, the Northwest Territories and Nunavut.

British Columbia has published all its orders-in-council from 1872 onward, but our team has decided to only target data back until 1990 to match the federal source.

Data collection

The IJF team collects all data from the sources detailed above, both through automated processes and manual scraping, depending on the update cycle of each respective data source.

The appointments database’s goal is to collect all appointments publicly released through orders-in-council. Data sources hosting PDF documents are collected with the original file intact. All our data is hosted on Amazon S3.

Extraction from unstructured text using LLMs

Our data pipeline parses unstructured data (such as arbitrary or document embedded texts) with the use of a GPT wrapper. We input into an Azure GPT-4o endpoint just the raw text and a prompt to extract and label important data points.

Schematization

After the data is run through the Azure GPT-4o LLM, the output is mostly structured to begin with. We match the AI output to a predetermined schema and make any necessary adjustments to have the former fit the latter. These are mostly data type policies to conform with our backend database.

Data cleaning

It is expected that the GPT-4o generated data is not bulletproof. The input data is almost always dirty to begin with (unstructured sentences, OCR from scanned documents, etc.) and consequently you can expect the output to inherit some of those attributes. 

Many instances of data cleaning involve culling the dataset of invalid dates, duplicates and partial entries. Moreover, our team has decided to prune specific terminology for legibility. For example, judges in British Columbia could be appointed to the title of “the honourable judge” or “hon. judge,” which can both be simplified to a title of “judge.”

Data for Canadian democracy

The IJF’s databases turn public records into public power. Explore millions of entries on lobbying, donations, contracts, access to information releases and more — and uncover the stories hidden in the data.