Guidance

Analysis function reproducible analytical pipelines (RAP) strategy 2023: DESNZ implementation plan

Updated 18 July 2023

Introduction

As part of the Analysis Function Reproducible Analytical Pipelines (RAP) strategy, public sector organisations must publish how they will meet the goals of the strategy. This document sets out how the Department for Energy Security and Net Zero (hereafter the Department for Energy) will work to implement the Analysis Function RAP Strategy during the 2023 calendar year. We set out how we will respond to the requirements of the strategy to provide the right tools, the right capability and the right culture.

Making 2023 a successful year for the Department for Energy

The Department for Energy was formed in February 2023 out of the Department for Business, Energy and Industrial Strategy (BEIS). As such the genesis and groundwork for this strategy was developed within BEIS. The adoption of RAP was in its relative infancy in BEIS and skills were spread across the large analytical community. This is also the case in the Department for Energy and whilst there are pockets of activity, there is currently no dedicated RAP resource or central team to work with business areas to transform existing processes. The establishment of the new department will also have some impacts on the speed of adoption of RAP and the actions identified in this workplan.

Nevertheless, Department for Energy recognises the potential benefits of RAP. Senior analytical leaders have endorsed adoption of RAP principles and committed to:

• increasing understanding and awareness of RAP by establishing a centralised internal information repository by summer 2023 - this will bring together existing information and links to resources from other organisations on RAP, such as how to plan and develop RAPs, training pathways and case studies
• supplementing the new internal information repository by developing specific products to help Department for Energy analysts develop RAPs by summer 2023 – for example a set of RAP prioritisation criteria which draws on the approach published by the Office for National Statistics (ONS) and guidance on the tools available on analytical platforms
• supporting senior leaders to work with teams to review workflows and identify RAP projects that should be carried out in 2023 or over the longer term – we plan to do this work by autumn 2023
• reviewing the existing internal analytical guidance and assurance processes, for example the Quality Assurance College, to identify where RAP principles can be incorporated
• establishing a Department for Energy RAP Support Network by summer 2023. – this will be a self-forming network, affiliated with the established Coffee and Coding Network, that will provide peer support and advice on RAP implementation and best practice
• communicating this workplan and raising the profile of RAP through presentations at internal meetings and RAP training for senior managers

Local strategic plan

The right tools: summary at the start of 2023

We have carried out an assessment of the department’s IT tools to support production of RAPs. The Cloud Based Analytical System (CBAS) ^{[footnote 1]} provides all the tools needed for analysts to produce minimal viable RAPs. Through CBAS, analysts have access to:

• version control software in Gitlab
• open source programming languages of Python and R with integrated development environments such as PyCharm and R Studio
• tools for managing environment and packages such as pipenv and renv
• libraries and packages, such as PyPl

Every user on CBAS has a home drive, shared file, and database storage. CBAS is managed by a dedicated team within the Analysis Directorate which enables flexible development of tools needed to support analytical work in the department.

The department’s main corporate IT environment, Cirrus, has some RAP tools, including R Studio and Power BI. But these tools can be limited in their usefulness because of a lack of Git version control, virtual environments, and fileserver and database storage. This means that Cirrus is not the preferred platform for developing RAP workflows. The department is also developing a data collection tool that, in future, could support RAPs, or components of RAP. But this functionality is not yet suitable for end-to-end analytical RAPs.

The right capability: summary at the start of 2023

Some parts of the department have the skills to use RAP already. Some analytical teams have started to develop analytical workflows in line with RAP principles. For example, the Energy Statistics team are transforming predominantly Excel based statistical production processes into reproducible pipelines using Python, SQL, and Git. Some automated processes have been developed for producing our Energy and Emissions Projections. For example, in some cases R Markdown is being used for reports and enhanced quality assurance using a R Shiny comparison tool. There are other areas of activity and interest across the analytical community in Department for Energy.

There are successful existing frameworks for technical knowledge sharing amongst analysts, such as:

• our in-house code training offer for R and Git
• Coffee and Coding
• the Data Science Project Mentoring Scheme

Specific learning pathways for RAP are not currently advertised within Department for Energy. Department for Energy is a member of the Analysis Function RAP Champions network, but there is no centralised team in Department for Energy with responsibility for developing RAPs or supporting teams in doing so.

We will build on existing areas of knowledge, experience, and activity to increase the collective capability of analysts to produce RAP workflows and transform existing processes, where there are clear benefits in doing so.

The right culture: summary at the start of 2023

The Department for Energy Analytical Leadership Team are aware of the potential opportunities and benefits of RAP. Teams are supported in developing their technical skills and in developing RAPs. But practical experience of working with RAP is mixed and greater understanding of the pathway to RAP is needed.

A Working Group was formed in summer 2022 from analytical teams across the department to produce our Implementation Plan. The Working Group was led by the Head of Energy Statistics, working closely with the Head of Profession for Statistics and Head of Advanced Analytics. The initial steps to implement this plan will be taken forward by this Working Group, under the mandate of the Director of Analysis and Analytical Leadership Team.

Appendix A: Detailed implementation plan

Tools

Action	2023 activities	Status	Success criteria / metrics
Owners of analytical platforms will give analysts access to the right tools	We will maintain access and tools for RAP on CBAS. We will not mandate use of a specific programming language for RAP development. Instead we will allow analysts flexibility to choose from languages such as R, Python, or SQL according to need and skills.	This work has been completed.	We will have achieved success when all new code projects use Python, R or SQL, unless there are exceptional reasons not to.
Owners of analytical platforms will give analysts access to the right tools	We will continue to work on connecting the Digital Data Management System (DMS) platform to CBAS to facilitate end to end RAP workflows across different platforms.	This work is in progress.	We will have achieved success when DMS and CBAS can share data with minimal delay, maintaining a single source of truth.
We will assess ways to provide open source access to our code	CBAS does not currently allow git repositories to be shared outside of that environment. This means we will concentrate on internal sharing across CBAS connected organisations for open Git repositories, code libraries, and shared packages.	This work is in progress.	We will have achieved success when at least three examples of shared code are being used across team boundaries within Department for Energy by the end of 2023. We will continue progress towards Open CBAS during 2023.

Capability

Action	2023 activities	Status	Success criteria / metrics
Analyst leaders will ensure their analysts build RAP learning and development time into their work plans	We will provide increased visibility of the RAP training pathways through links and information on the centralised information repository.	This work is in progress.	We will measure success by: – the number of hits on the central information repository, when it is established – noting an increased demand for RAP training.
Analyst leaders will help their teams to work with Digital, Data and Technology (DDAT) professionals to share knowledge.	We will encourage multi-disciplinary knowledge sharing across professions through existing structures such as Coffee and Coding and establishing a Department for Energy RAP Support Network.	This work is in progress.	We will have achieved success when we have established a Department for Energy RAP Support Network, which we plan to do by the end of summer 2023.
Analyst managers will build extra time into projects to adopt new skills and practices where appropriate.	We will provide advice on what sort of time investment and skills are needed across different types and sizes of RAP project. This will guide managers in their project planning.	This work is in progress.	We will have achieved success when we complete and share guidance on RAP project planning.
Analyst managers will learn the skills they need to manage software.	We will consider and include RAP skills in learning and development plans for analysts. We will do this through discussion with their managers.	This work is in progress.	This work is ongoing.
Analysts will learn the skills they need to use RAP principles.	We will: – support analysts to increase their knowledge, understanding and access to RAP skills through a centralised information resource on the internal Analysts Hub – consider and include RAP skills in learning and development plans for analysts – we will do this through discussion with their managers – use the existing structure of Coffee and Coding sessions to share knowledge on RAP projects, tools and best practice – establish a Department for Energy RAP Support Network, affiliated to the Coffee and Coding network – this will provide peer support and a forum for those planning and undertaking RAP projects	This work is in progress.	We will measure success by: – the number of hits on the central information repository, when it is established – encouraging analysts at Department for Energy to complete the “Coding in analysis and research” survey and assessing the extent of RAP skills and uptake – holding at least 2 Coffee and Coding sessions that relate to RAP in 2023/4 – beginning to set up a Department for Energy RAP Network by summer 2023

Culture

Action	2023 activities	Status	Success criteria / metrics
Department for Energy will choose leaders responsible for promoting RAP and monitoring progress towards this strategy within organisations	Director of Analysis along with the Department for Energy Analysis Leadership Team endorses this RAP Implementation Plan.	Completed	The Implementation Plan is complete. Further activity is planned for the first half of 2023 to put in place documentation and guidance for teams. We will monitor progress at least annually in the Analytical Leadership Team (ALT) meeting.
Department for Energy will form multidisciplinary teams that have the skills to make great analytical products. Some members of these teams may specialise in developing analysis as software.	We will include RAP skills in recruitment activity, where appropriate.	Not started	We will have achieved success when job adverts include RAP skills, such as coding.
Department for Energy analyst leaders will encourage a “RAP by default” approach for all appropriate analysis.	The Analytical Leadership Team have agreed to consider applying RAP methods and principles in analysis.	This work is in progress.
Department for Energy analyst leaders will encourage a “RAP by default” approach for all appropriate analysis.	Department for Energy analyst leaders will be supported to encourage this approach by an increased understanding of RAP through ‘introduction to RAP’ training for senior leaders.	This work is in progress.	We will have achieved success when we start training in 2023. Training will involve using Analysis Function resources or courses if they are available.
Department for Energy analyst leaders will lead their RAP champions to advise analysis teams on how to use RAP.	We will establish a Department for Energy RAP Support Network of those undertaking or planning RAP projects to provide peer support, guidance, and advice.	Not started	This work will start by summer 2023.
Department for Energy analyst leaders will help teams to incorporate RAP development into their workplans.	ALT will work with their teams to review and identify workflows and projects that will benefit most from RAP. They will follow training and RAP criteria to do this.	Not started	This will be done by autumn 2023.
Department for Energy analyst leaders will identify the most valuable projects by looking at how much capability the team already has and how risky and time-consuming the existing process is.	We will develop a set of Department for Energy RAP criteria for projects, based around the ONS RAP project criteria.	This work is in progress.	These criteria will be shared internally by summer 2023.
Department for Energy RAP champions will support leaders in their organisation to achieve the aims of this strategy by acting as mentors, advocates, and reviewers.	Department for Energy RAP Champions will continue to attend x-gov RAP Champions Network and disseminate discussion and actions.	This work is in progress.	This project is ongoing. We are continually learning and sharing of good practice from other organisations.
Department for Energy RAP champions will manage peer review schemes in their organisation to facilitate mutual learning and quality assurance.	We will establish a Department for Energy RAP Support Network of those undertaking or planning RAP projects to provide peer support, guidance, and advice.	Not started	This work will start by summer 2023.
Department for Energy Analyst managers will evaluate RAP projects within organisations to understand and demonstrate the benefits of RAP.	We will develop a framework for evaluation of RAP projects following a review of workflows to identify RAP projects.	Not started
Department for Energy Analyst managers will instruct their teams to use RAP principles whenever possible.	We will consider RAP methods for all projects, as agreed by ALT.	This work is in progress.
Department for Energy analysts will engage with users of their analysis to demonstrate the value of RAP principles and build motivation for development.	We will compile some case studies for RAP projects within Department for Energy to show the benefits of RAP.	Not started
Department for Energy analysts will produce their analysis using RAP.	We will continue to support RAP projects underway in Department for Energy. We will identify new RAP projects through the ALT review of workflows.	This work is in progress.

Appendix B: Assessment of tools at Department for Energy, December 2022

For Reproducible Analytical Pipelines that meet the minimum criteria	Comment	Availability for use by analysts
Version control software, that is, git	Gitlab is available on CBAS	Available
Oopen-source programming languages and flexibility to add more – Python, R, Julia, JavaScript, C++, Java/Scala and so on	Department for Energy Analysts have access to Python and R on CBAS. SQL, although not open source, is also available and well supported.	Available
Package and environment managers for each of the available languages	Python and R have tools for managing environments and packages on CBAS (e.g. pip and renv)	Available
Packages and libraries for open-source programming languages, either through direct access to well-known libraries, for example, npm, PyPI, CRAN, or through a proxy repository system, for example, Artifactory	Access to PyPI packages through CBAS. Mirror of CRAN.	Available
Individual storage, for example, home directory	Every user on CBAS has a Home drive	Available
Shared storage, for example, s3, cloud storage, with fine-grained access control, accessible programmatically	CBAS provides access to MS SQL Server, File Share, PostgreSQL, MS Access	Available
Integrated development environments suitable for the available languages – RStudio for R, Visual Studio Code for Python and so on	Rstudio for R, PyCharm for Python. MS Code	Available
For further development:
Source control platforms, for example, GitHub, GitLab or BitBucket	Department for Energy Analysts have access to Gitlab on CBAS	Available
Continuous integration tools, for example, GitHub Actions, GitLab CI, Travis CI, Jenkins, Concourse	Not currently available	Not currently available
Make-like tools for reproducible workflows, for example, make	Not currently available	Not currently available
Relational database management software, for example, PostgreSQL, that is available to users	PostgreSQL is available, although not widely used.	Available
Orchestration systems for pipelines and workflows, for example, airflow, NiFi	Not currently available	Not currently available
Internal-facing servers to host html-rendered documentation	Able to internally host R Shiny apps.	Partially available
External-facing servers with authentication to host end-products such as web applications or APIs	Able to share R Shiny apps to CBAS light users	Partially available
Big data tool, for example, Presto or Athena, Spark, dask and so on, or access to large memory capability	Not currently available and minimal demand to date	Not currently available
Reproducible infrastructure and containers, for example, docker	Not currently available	Not currently available

1. CBAS was set up in 2015 and provides services to over 600 users across 19 organisations (as at February 2023). It is a secure, collaborative platform, up to OFFICIAL-SENSITIVE. CBAS allows the easy share of data, within and between teams, and across organisations. It hosts many software tools, both opensource and commercial, and has web application functionality to host user-created dashboards and apps. CBAS has a dedicated support team of data management and IT specialists, located within Department for Energy. ↩