Recognizing Data Science Skills
with BDV Data Science
Badges and Training Labels
Nik Swoboda
BDVe Project and the Universidad Politécnica de Madrid
Webinar on Big Data Skills Accreditation – 17/09/2020

Overview
Today I will present work which is the result of a collaboration between
the BDVe project and the Skills and Education Task Force (TF9) of the
BDVA.
The goal of this work is to develop frameworks for the recognition of
skills in data science which address the principal needs of the main
stakeholders.
2

Outline
 Discuss Stakeholder’s Needs
 BDV Data Science Badges for Formal Education
o The key aspects of the proposal
o The requirements and application process
o Issuing and displaying badges
 BDV Data Science Training Labels for Non-formal Education
o Motivation
o The key aspects of the proposal
o The current state of the proposal
 Summary and Future Work
3

Stakeholder’s Needs:
Data Scientists
 Credentials which are:
owidely recognized (DS-N1)
oeasily verified online (DS-N2)
 A simple way to digitally display their skills
online and in social networks (DS-N3)
 Mechanisms to recognize skills acquired
through informal and non-formal training.
(DS-N4)
44

Stakeholder’s Needs:
Employers of Data Scientists
 Candidates with credentials which are both
granular and individual (EM-N1)
 Tools to verify the authenticity of credentials
(EM-N2)
 A framework which simplifies the comparison
of skills throughout the EU (EM-N3)
 Influence in the design of the training data
scientists receive (EM-N4)
 A scheme which can quickly adapt to changes
in the data science ecosystem (EM-N5)
55

Stakeholder’s Needs:
Educators Who Train Data
Scientists
 Publicity for their programs and the value
of a branded recognition of their programs
(ED-N1)
 Recognitions for the partial completion of
their programs (for students who want to
work while studying) (ED-N2)
 A mechanism to clarify the changing
needs of industry and recommendations
regarding how to adapt to those needs.
(ED-N3)
66

BDV Data Science Badges for Formal
Education
 The importance of formal education is widely recognized and
respected
 The goal of the proposal is to:
provide added value to the existing offer of formal education in data science by
trying to address the previously discussed stakeholder’s needs
7

Key Aspects of the Badge Proposal
 Use Open Badges (DS-N2, DS-N3, EM-N1, EM-N2, ED-N2)
 Experts from industry and academia will contribute to both defining and
maintaining the badge scheme (EM-N3, EM-N4, EM-N5, ED-N1, ED-N3
 Badges will only be issued by trusted third parties (DS-N1, ED-N1)
If the badges become popular, their requirements will influence the overall
training of data scientists (DS-N1, EM-N4, ED-N1)
*Note: DS-N4, related to non-formal and informal learning, will be
addressed by the second recognition system.
8

BDV Data Science Badges (for Formal Education)
Design
IssueDisplayVerifiable
Useful Metadata
9

10
BDV Data Science Analytics Badge v1.0 – Academic Level
Required Skills:
DSA.1. Identify existing requirements to choose and execute the most appropriate
data discovery techniques to solve a problem depending on the nature of the data
and the goals to be achieved.
DSA.2. Select the most appropriate techniques to understand and prepare data
prior to modeling to deliver insights.
DSA.3. Assess, adapt, and combine data sources to improve analytics.
DSA.4. Use the most appropriate metrics to evaluate and validate results, proposing
new metrics for new applications if required.
DSA.5. Design and evaluate analysis tools to discover new relations in order to
improve decision-making.
DSA.6. Use visualization techniques to improve the presentation of the results of a
data science project in any of its phases.

Applications to Issue Badges
 Each university degree program
interested in issuing badges must
prepare an application. Applications
include the following:
o General description of the program
o For each required skill:
• Time dedicated to learning the skill
• Evidences used to confirm the acquirement of
the skill
11

Issuing Badges to Students
 The student learns and demonstrates the acquisition of all the skills
required by the badge
 Using the Badge Platform: The student applies for the badge,
attaches an evidence file and accepts the platform’s data protection
agreement
 The university reviews the application
 Using the Badge Platform: The university issues a badge to the
student
12
Awards
Badge

Displaying Badges
 Students receive an email with a link to the badge
 Students can upload their badges to any Open Badge Repository
(Backpack), include them in digital documents (like CVs), and share
them in social media
 Viewers of badges can
o Verify the authenticity of the badge
o Access the metadata contained in the badge
• Who issued the badge, required skills, date issued, etc.
o Open linked evidence files
13
Displays
Badge

BDV Data Science Training Labels for Non-
formal Education
The importance of non-formal education* in data science is growing:
 In this quickly changing field, professionals constantly need to retrain
 With the growing demand and the insufficient supply of data scientists
o Employers can offer training to existing employees to meet needs
o Employees interested in career changes can retrain
 Working professionals can find offerings in formal education impractical
 Non-formal training can be cost effective
Goal: Propose a second skills recognition program for non-formal training.
15
* Education that is institutionalized, intentional and planned by an education provider. The
defining characteristic of non-formal education is that it is an addition, alternative and/or a
complement to formal education within the process of the lifelong learning of individuals.
(UNESCO)

Key Aspects of the Label Proposal
 Transparency - Establish a common set of essential data useful for all (DS-N4, EM-
N3)
 Ease Comparison - Define a common format to simplify side-by-side comparisons
(DS-N1, DS-N3, EM-N3, ED-N1)
 Unify concepts - Provide a common set of core data science skills (ED-N3)
 Adaptability - Program coordinated by both academics and professionals (EM-N4,
EM-N5, ED-N3)
 (Attempt to address some of the specific needs of non-formal training …)
* The other needs (DS-N2, EM-N1, EM-N2, ED-N2) must be addressed by the
individual programs
16

Specific Challenges Related to Non-Formal
Training in Data Science
For students: How to pick from the overwhelming supply?
o Which programs are more highly valued by industry?
o What is the right level of training for their experience and expectations?
For industry: How to compare different types and quality of training?
oHow ‘serious’ are the programs: duration, level, testing, identity
verification
For education providers: How to stand out from the rest?
o Clearly communicate their offer, attract students, ensure the quality of their
training, …
17

Choosing Food is Made Easier with Standardized
Nutritional Information Labeling
18

Criteria for the BDV Data Science Training
Label
 Basic information
o Name of the course, name of the training provider, type of provider (university, not for profit
organization, … )
o Cost
o Language
 Content
o Audience (technical, non-technical)
o Is there a focus on a specific domain?
o Total hours and breakdown dedicated to:
• Business Analysis, Data Preparation, Model Generation, Model Validation, Visualization
o Kind of training (hours of online interactive classes, in person classes, …)
 Kind of testing
19

A Draft of the BDV Data Science Training Label
20

Summary and Future Work
The BDVe project has produced:
 Two ready to be implemented frameworks for recognizing skills in data
science
o BDV Data Science Badges with a focus on formal education (piloted)
o BDV Data Science Training Labels with a focus on non-formal education (soon to be
piloted)
More Information: https://www.big-data-value.eu/skills/skills-recognition-
program/
What remains:
 Obtain support and branding for the programs with more influence and
wider recognition
21

Feedback?
Contacts:
 Ernestina Menasalvas (ernestina.menasalvas@upm.es)
 Nik Swoboda (nswoboda@fi.upm.es)
 Ana M. Moreno (ammoreno@fi.upm.es)
22