8  Professional Competence

Standards and Breadth of Knowledge

8.1 Skill Areas

The Alliance for Data Science Professionals defines the standards for data science practitioners in a table outlining five skill areas (A–E), with ethics being a sixth category included in each of the other five. See Appendix 1 for details.

8.2 Breadth of Knowledge Requirements

Certification Requirement
DSP Demonstrate applied understanding and knowledge in three of the five skill areas. They may have limited understanding in the remaining two areas.
AdvDSP Demonstrate deep understanding and knowledge in Skill Area E and at least two other skill areas; display appropriate general knowledge in the remaining two skill areas.

8.3 Proficiency Levels

The following table describes the levels of knowledge used to assess applicants:

Skill Level Proficiency Experience
Limited Limited conceptual knowledge Minimal experience—read about it, some education and some practice with tools; some exposure in educational training setting
General General conceptual knowledge; theoretical knowledge applied in education or training context Completed formal education, including capstone project; has factual, procedural, and theoretical knowledge; can interpret and evaluate relevant information; is aware of different perspectives
Applied Applied knowledge Performs with supervision or mentoring
Deep In-depth knowledge Proficient with state-of-the-art approaches without supervision; advanced practical, conceptual, or technological knowledge and understanding of the field of work, enabling the applicant to create ways forward in contexts in which there are many interacting factors; understands different theoretical perspectives; can critically analyze complex information
Expert Expert knowledge Advances the state-of-the-art

8.4 Ethics and Efficacy

It is important that all professionals working within the field of data science have a clear understanding of the ethics that underpin the collection, management, use, and communication of the data and the results with which they work. It is equally important that a data scientist takes responsibility for the assurance of the models they build. Assurance covers both the efficacy of the application and the ethical nature of its design and implementation.

These attributes are not something that can, or should, be assessed as one standalone criterion. Rather, when completing this application, you should wherever possible include reference to your knowledge and working practices relating to the appropriate ethical considerations such as the following:

Data considerations:

  • Collection, validity for use in the intended purpose, permission for usage, storage, security

Model considerations:

  • Development, testing (e.g., fairness, bias, error rates), usage (how the model and results could be used for an unintended purpose?), and transparency

Communication considerations:

  • Explanation of why the science is required, the results achieved, and how misinterpretation of the results can be minimized

Legal and regulatory:

  • Relevant laws and permissions of usage for data (including legal rights of individuals, privacy, and anonymity)

Efficacy considerations:

  • Quality assurance of code and data
  • Validation of model fit
  • Robustness of the model and software implementation
  • Ongoing monitoring of model implementation
Note

The list above is not exhaustive. It serves as a guide to help applicants show the DSCC they are aware of the professional expectations of those who work in this field. Applicants should include any other areas of ethical and efficacy considerations they feel are important within their area of expertise.

8.5 Confidentiality and Conflicts of Interest

Applicants may alert the DSCC that some members of the committee should not see their application materials due to confidentiality, conflict of interest, or other specified reasons. Members of the DSCC may also recuse themselves from reviewing certain applications for such reasons (see Section 17.1).