CODEx Project: Alpha Phase insight
The FRC launched the Company & Organisational Data Explorer (CODEx) Project in September 2023. The project is supported by Companies House, FCA, the Charity Commission, HMRC and others and is funded by DSIT as part of the Regulators’ Pioneer Fund*.
The project is designed to optimise the value for the public and for regulators from the millions of annual reports that are filed in a structured data file format (iXBRL) in the UK. These annual reports are rich in accounting, financial and reporting information, by making the data easy to use and reuse we can help support growth and economic decision making.
We are accelerating the use of XBRL data from the iXBRL annual reports, by:
- building an iXBRL viewer (the Viewer) – a public tool to display individual or a small number of iXBRL annual reports, showing the embedded financial and non- financial tagged data in the context of the human-readable report; and
- developing skills, blueprints, and private tools for XBRL bulk data analysis by regulators and agencies (the Toolkit).
After completing the Discovery phase, the CODEx project moved into the Alpha phase. Alpha is about building prototypes, testing different approaches and considering the best way forward.
Developing a prototype for the Viewer
For the Viewer, Alpha has meant focusing on understanding the core user journey. We wanted to evaluate how users would likely interact with the Viewer and see how they responded to different design options. We focused our research on those searching for iXBRL reports from Companies House and FCA’s publicly available iXBRL repositories, opening them in the Viewer and viewing facts.
We designed high fidelity prototypes of the Viewer based on the GOV.UK Design System and tested it with users. A total of 14 interviews were conducted with the aim of iteratively testing the designed prototype. Interviewees were presented with options of how the Viewer core user journey would work and their responses were reflected in the refinement of the designed prototype. As developing the Viewer to support users with accessibility needs is a key driver, we made sure we also involved users with accessibility needs.
We learned a lot from the exercise and overall, we were pleased that testers understood the value of the tool and how it could improve their use of this data.
Based on the results of the Alpha phase, we have concluded that it is possible to develop a solution that complies with the Technology Code of Practice and adds value to users. In addition, based on the lessons learned, we have reconsidered how best to deliver a solution that meets as many of the user requirements identified in the Discovery phase. We are taking these lessons through to the next stage of Viewer development.
Building the foundations for a Toolkit
Using the Viewer is a great way of looking at specific data points within an annual report, but we also want to build out our capacity to use the data en-mass to support our understanding of trends, identify quality issues and provide analysis that supports policy. To do this we have been working on a regulatory Toolkit.
To build a Toolkit that makes use of structured accounting data (XBRL) you must first collect the data (from public sources) and transform it (parse) into tables that can be used in a database. We set up proof-of-concepts for parsing using commercial and open-source XBRL-certified tools. This showed us that the open-source parser was useful but needed significant amounts of technical expertise to use. We shared our findings with other regulators and agencies in a show-and-tell session. The code we developed for the open-source proof-of-concept is available on GitHub.
Developing the Toolkit
Taking what we learned from Alpha we worked together with our development partners on a Minimum Viable Product for the Toolkit. To do this we ingested the iXBRL data from all available FCA filings and a sample of Companies House and charitable organisation filings into a relational database. We also ingested reference data, including industry classification, stock index constituents, foreign exchange rates and country of origin. We turned the resulting tables into a data model that connected the data together in a way that allows us to analyse and use the data to support our key use cases.
Working together
To test out the value of the data model and toolkit we organised two two-day workshops (‘data sprints’) in London and Birmingham. Potential users across our regulatory partners and agencies, including the FRC, the FCA, Companies House and the Charity Commission worked together, with the aim to:
- Explore specifications for tools and dashboards that meet regulatory or agency use cases.
- Build skills around working with the XBRL data.
- Inspire and engage participants to use XBRL data and experiment with it.
The ‘sprints’ were useful and proved our hypothesis, that there is significant value in the data once it has been processed into a database. The event also gave us lots of areas to focus on for the next development stage.
Find out more about CODEx or get in contact with the CODEx team.
*In November 2022, BEIS awarded the FRC a grant of £796,000 via the Regulators’ Pioneer Fund to deliver CODEx. The RPF is a grant-based fund to enable UK regulators and local authorities to help create a UK regulatory environment that encourages business innovation and investment. The current £12m round is being delivered by the DSIT.