European Data Science Academy

European Data Science Academy

Chris Phethean edsa-project.eu

University of Southampton [email protected]

What is EDSA about

ESTABLISH A EUROPEAN COMMUNITY OF EDUCATORS, LEARNERS, PRACTITIONERS, AND POLICY MAKERS ON DATA SCIENCE

Who we are

How do we deliver a syllabus to allow anyone to master Data Science?

Image: http://mappr.it/2016/02/24/data-science-for-beginners/

What is EDSA?

•  Surveys •  Interviews •  Dashboards

Landscaping

•  Modular, media-rich, multiple languages

•  Core, domain-specific, and technology-specific topics

Curriculum and

courseware •  Video lectures •  Professional training •  MOOCs •  eBooks

Courses and learning analytics

Initial Version of Module

Final Module: eBook & Course

Webinar – Final Recording

Webinar – First Recording

Collection of Module Materials Slides – First Version

Slides – Final Version

Stakeholder Communities Industrial Advisory Board

Demand Analysis Reflection

Module Dissemination Online & F2F

Learning Analytics

Curricula EDSA Analytics

Dashboard

Reconfiguring & Repurposing

Learners Learning Delivery

End to End approach

1. What is the current demand for data skills in Europe?

2. What training should be offered in order to accommodate this demand?

•  Surveys •  Interviews •  Dashboards

Landscaping data science

demand

Demand Analysis

The European Data Science Landscape

28 EU member states

19 Eurostat defined industry sectors

Sole traders to large

companies

584 survey responses

108 interviews

What is the current demand for data skills in Europe?

Organisation Eastern Europe

Northern Europe

Southern Europe

Western Europe

NA Total

Large 69 112 100 105 1 387

SME 36 75 81 45 4 241

Micro 5 14 7 9 1 36

Self-employed

2 4 5 3 0 14

NA 0 2 2 10 0 14

Survey & Interview Responses

Current Demand for Data Skills •  Importance of skills: How would you rate the

following skills for a data scientist?

Managers

Practitioners

•  How proficient are you (practitioners) or your teams (managers) in the following skills areas? (1=very poor; 5=very good)

What tools, technologies or languages should be covered on data science courses?

What other skills are needed?

0

5

10

15

20

25

Communication and presentation skills

Industry and business domain

knowledge

Teamwork Data management Social skills

% of interviewees

“The data scientist must be willing to adapt quickly in order to keep up.”

Data scientists must have the “soft skills required of interacting with businesses and guiding the people responsible for making decisions“

What training methods would you prefer for data science training?

Preliminary Conclusions •  Communication and persuasion skills

•  Social skills

•  Basic data literacy training for organisations

•  Skills analytics framework

•  Sector-specific courses

Challenges to finding training •  Majority of interviewees faced challenges with:

–  finding courses: “It is difficult to find courses also because there is no single platform to search within.”

–  gathering sufficiently detailed information on contents and course quality

–  finding tailored solutions aligned with team’s needs

•  Understanding one’s own status and demand (both as an individual and organisation) is an additional challenge: –  “The difficulty is finding or understanding where the

individual is in terms of his own skill-set and where he needs to develop, and then finding the resources to plug into those gaps.“

dashboard.edsa-project.eu

Dashboard Data: Job Demand in the

EU •  ~300K job posts from

–  Adzuna (UK, DE, FR, N) via API

–  Trovit (Additional 18 countries) via scraping

•  Extracting –  Time –  Geolocation –  Skills required –  Company

Country No. of Postings Extracted Austria 4,694 Belgium 10,764 Bulgaria 44 Czech Republic 6,232 Denmark 7,026 Estonia 2 France 34,881 Germany 20,938 Hungary 7,225 Ireland 18,919 Italy 11,543 Malta 38 The Netherlands 9,571 Norway 33 Poland 18,159 Portugal 21,784 Romania 17,946 Slovakia 3 Spain 9,210 Sweden 13,870 Switzerland 14,616 United Kingdom 89,343

TOTAL 316,841

The EDSA Curriculum

Foundations •  Foundations of

Data Science •  Foundations of

Big Data •  Statistical /

Mathematical Foundations

•  Programming / Computational Thinking (R and Python)

Storage and Processing •  Data

Management and Curation

•  Big Data Architecture

•  Distributed Computing

•  Stream Processing

Analysis •  Essentials of

Data Analytics and Machine Learning

•  Big Data Analytics

•  Process Mining

Interpretation and Use •  Data Visualisation •  Visual Analytics •  Finding Stories in

Open Data •  Data Exploitation

Addressing Demand EDSA courses portal: edsa-project.eu/resources/courses/

Curriculum Dimensions •  Courses to be tagged with pre and post skills •  Learning pathways aimed at:

–  Statisticians –  Analysts –  Managers, product owners, CEOs –  Programmers, developers and system engineers –  Data managers (incl. security experts)

•  Further dimensions: –  Tools and programming languages –  Type of data –  Industry sector –  Level

•  Basic, advanced, expert

Summary •  Challenges

–  Data Science is a broad, multidisciplinary subject that requires a mix of skillsets

–  How do we equip practitioners and their managers with the ability to assess their own skill profiles and identify gaps?

–  Tools and tech used in industry change rapidly, faster than academia typically updates courses

–  How do data scientists locate appropriate courses to best enable them to learn new skills?

–  How should courses be delivered in order to target the diverse range of backgrounds and skillsets interested in data science?

Data Science @ Southampton