Reproducible Analytical Pipelines (RAP) are the application of good practices from software engineering in analysis and research. They make analysis and research more reproducible, better quality, more trustworthy, and efficient. The recent Office for Stats Regulation review of RAP says "there is a huge potential to apply RAP principles both in and beyond official statistics".

RAP is the linchpin of the digital transformation of government analysis. Through the high quality automation it brings, more effort can be directed towards building the research products that our users need. Recent successes like the UK Coronavirus Dashboard show how powerful government data can be in answering the needs of the public when sound RAP principles are applied.

However, learning the practices in RAP represent a substantial challenge for many government analysts - especially in a context where there is little prioritisation to build this capability and where the tools needed can be difficult to acquire. Our Coding in Analysis and Research survey shows that although most analysts feel it's important to implement the RAP principles in their work, many fewer are able to do so.

Our approach to RAP is multi-pronged to tackle the cultural, capability, and technical barriers that exist:

• We operate a Champions network that engages across government and delivers mentoring, training, and advocacy within their respective departments
• We deliver consultancy and mentorship for organisations beginning their journey
• We advocate for the digital transformation of analysis with RAP at senior levels in government
• We monitor progress towards RAP and evaluate interventions,
• We role model good practice in our own analytical work

Through our research of coding in analysis across government and the research of colleagues we identify next steps for public sector organisations and advocate for the inclusion of RAP principles at the heart of government analytical strategy.

Speaker: Alexander Newton, ONS