Skip to content

Latest commit

 

History

History
73 lines (54 loc) · 3.25 KB

governance.md

File metadata and controls

73 lines (54 loc) · 3.25 KB
Raw

THREE TIER GOVERNANCE

Created by @eric_bickel

In order to maintain quality of work, it is incumbent upon the organization to ensure proper governance. We maintain quality of work by embedding governance at every step of the process: data ingestion to data analytics to report generation.

To encourage governance that is molded to each team, each team must appoint individuals to monitor and track each of the three main processes: DATA, ANALYTICS and REPORTING. This document aims to provide guidance for those individuals that are responsible for each of these three processes.

DATA GOVERNANCE

Great analysis must begin with great data. For credibility and legality, the data that we use must be properly sourced and adequately cited throughout our processes.

Additionally, data cleansing efforts must be properly defined. When in-depth details are not necessary, a brief summary or guideline of the process(es) used should be provided. This serves two purposes:

  • Ensure that we are using best practices for data cleansing to avoid reducing the quality of the incoming data.
  • Help downstream analysis by clearly defining how the data has been processed. This may be relevant to determining which analytic methodology to use.

To implement these data cleansing guidelines, dedicate a section called "Data Ingestion" within each team's README file. List all data sources along with links to their data.world location. Include descriptions of any cleansing that has been done to the data in the Project.

ANALYTICS GOVERNANCE

We recognize that analysis is both an art and a science, so flexibility and creativity are absolutely encouraged. It is also necessary to ensure that the chosen methodology is appropriate and correctly implemented.

Dedicate one section to analysis Within the team's README file called "Data Analytics". This section can be either:

  • A written section outlining the analysis performed on the data
  • A link to the analysis performed (e.g. ipynb or rmd files hosted on the team’s GitHub repo)

It is the responsibility of the individual Project Leads to formalize a process for managing changes to the analytic methodology as defined in the README. A formalized change process is especially relevant when proposed changes materially affect the analysis. The Project Leads should designate individual(s) as reviewers of such requests to promote decision making transparency.

REPORTING GOVERNANCE

Effectively communicating the analytic results to the broader Data4Democracy community and the general public is essential. Even the best analytic results can lose potency when communication methods are not ideal.

Each team should appoint one individual to champion each report. There should be at least two editors for each report to review and ensure accuracy. One editor should be appointed that was substantially involved in data ingestion and another that was substantially involved in the analytics. Any additional editors can be appointed at the discretion of the Project Lead.

Each report must link back to the Project's README file to ensure total transparency.

If a Project materially deviates from these Reporting suggestions, add a section to the Project's README file under the heading "Data Reporting" and detail the reporting practices of the Project.