# TSG National Skills Subgroup Meeting, 12.04.22

###### tags: `focus session`

## Context and background

The purpose of the Turing Skills Group (TSG) is to expertise and advice from a representative group of the Turing's community to support the Research and Innovation Advisory Committee's (RIAC) Skills Working Group (SWG) to make recommendations to Boards, Committees, and the Turing Management Team, where appropriate.

The TSG subgroup for National Skills is led by Matt Forshaw and Bridget Nea.

## Purpose and goals

* To identify national priorities for skills, its a useful exercise to hash out the landscape of the key terms we're using, the audiences we're defining for, and take stock of the resources we have and are lacking.

* The focus questions below start with the basics as we try to define a common language when we discuss national skills and establish a shared lexicon in a landscape where terms can be slippery and take on different meanings contextually.

* Please consider what the Turing can/can't and should/shouldn't do in relation to the questions below. Please also consider what falls outside of the expertise/role of the Turing as we try to identify gaps and hard limits for our involvement in national skills.

* Other points for consideration to the questions below relate to the scope of our area of work, namely, is our area of focus too broad when we consider factors like audience?

* With your input and expertise, the outcome of this discussion will also inform the [Data Skills Taskforce](https://www.dataskillstaskforce.com/) (DST) focus areas and thematic structure for the coming year as we feedback on national skills.

* For context, the four current priority thematic areas for the DST are listed below and over the next six months, we aim to tease out overarching gaps in the landscape for a wide variety of stakeholders:

::::info
1. **Organisational Data Readiness** (AI Council Roadmap: “National, Cross-sector Adoption: Business as smart adopters”).
1. **Professionalisation** (NDS: 5.1.1 Definition of data skills and role descriptors, AI Council Roadmap: Recommendation 6: Commit to achieving AI and data literacy for everyone. Recommendation 5: Make diversity and inclusion a priority).
1. **School-level data literacy** (AI Council Roadmap: Recommendation 6: Commit to achieving AI and data literacy for everyone. Recommendation 5: Make diversity and inclusion a priority).
1. **Equality, Diversity and Inclusion** / Widening Participation(AI Council Roadmap: Make diversity and inclusion a priority.)
::::

## Key questions :question:

1. What do we mean by data skills on a national framework? What definitions and descriptions are we working from and using? E.g. data literacy and digital literacy often used interchangibly.

[*National Data Strategy, 5*](https://www.gov.uk/government/publications/uk-national-data-strategy/national-data-strategy#data-1-2)

2. Who and what are our target audience(s) when we discuss national skills?
- What sectors/areas should be our focus be when we define national skills? E.g. data science practicioners, primary and secondary level education, tertiary education, further education, the wider workforce.

1. What existing resources (policy, tools, research, groups, practices) can we draw from to inform our understanding of national skills?
*or* What kind of resources or outputs are missing/could be provided by the Turing's national skills projects? E.g. a consultation from the DST, a data/AI skills strategy.

## Collaborative notes :lower_left_ballpoint_pen:
* Themes and priority areas:

- Question 1
- **GC:** Across a wide range of areas, plotting an idea's journey is particularly difficult. Requirement in CS and Stats, and application in real world, three discrete and interconnected disciplines. Each have their own distinct pedagogy.
- How to explain someone at 16 what a journey to a DS would be
- A great deal is explored rather than taught.
- **RH:** It depends at what level you plan to define it to.
- It can almost become meaningless.
- Journey is not just towards Data Scientist.
- Always interested in Quant skills for geographers
- **PP:** Data skills can include quant methods without scaring people as much, and helping see, how is the DST responding to this?
- **MF:** lack of role descriptors, issues entering the workforce, slippery terms for 'data scientists'
- **NW:** challenge - setting training, mismatch in the needs of the workforce/employability vs. degree programmes
- **GC:** being able to handle large data sets/level of skill in world of coding & accessing open source solutions

- Question 2
- **BN:** Are we talking about primary/secondary educators, workforce, etc?
- **NW:** When teaches, likes to tier skills across 3 groups - citizen, some interaction with DS in job, and technical DS/ML roles. Everyone needs a core set of data literacy skills, esp. as data technologies become more widespread and have more impact on daily life. Think critically about sources of data, parsing data, see what's trustworthy.
- Direct tuition/online/in skills?
- **MF** cross-over between audiences (board room/school level), dialectic learning flows - initiatives/GC
-
- Question 3
- RH: Data science and AI - laylevel training/guidebook - what structure? Work w/ learned societies - not sure whether Turing should do this?
- NW: Intro to data science, materials cross over b/w audiences (year 8 + entering workforce + execs considering adopting data technologies or managing DS work)
- PP: Turing - getting people together to set industry standards is key, possibility of a handbook
- NW: Raspberry Pi's Big Book of Computing Pedagogy is a good example of a handbook for educators

### Main Challenges:
- NW: Lots of toy problems to teach data science.
- Huge drop in confidence when students are transitioned onto a real problem with real data.
- Haven't seen data collection, weird problems,
- oversimplificiation of the problem.
- Challenges on realistic data sets, many online sources like Kaggle reward tidy, 'well-behaved' data that's not representative of a lot of routine work.
- First problem in DS has kids collect their own data.
- Students collect data on school energy use, lighting and activity. Half term break occurs part-way through data collection. Many creative explanations for anomalous week, many fail to draw the connection to half-term.
- When you collect your own data, you have to decide where to sample, how long, how much data needed. Some companies really struggling to do this is it's their first foray into AI/ML projects.
- Trustworthiness of data/decision making/critical thinking around the problem and more overarching problems

- GC: Data science and AI being communicated across disciplines/non-cognate ]#

- MF: Opportunities/large scale survey of the youth re: critical assumptions