Case study

Open Policy Finder

Researchers need to find the publishing policy relevant to their journal or publisher. At Jisc, I worked on the search, documentation, and data quality that help them answer that question.

Context
Jisc / April 2023–February 2026
Focus
Search, documentation, Elasticsearch, data quality
Project
openpolicyfinder.jisc.ac.uk

People search with partial titles, acronyms, publisher names, and imperfect recall. A relevant result also depends on the source records being complete enough to use, while the guidance needs to stay editable by the content team.

AreaWhat I built and why
Search

Built Elasticsearch-backed search, ranking, filters, and result views.

Users can search beyond an exact title match and narrow the records they see.

Documentation

Built Markdown parsing and React rendering, connected to Directus publishing.

Content teams can maintain help and API guidance without editing the page implementation.

Data quality

Worked with the data team to identify and correct missing and malformed records before launch.

Source-data issues can be fed back to the people maintaining the records.

ChoiceWhy it matters
Support imperfect queriesFuzzy matching, ranking, and filters address the ways people actually search. An exact lookup alone would make users responsible for remembering the dataset’s naming conventions.
Separate content from its renderingMarkdown is parsed into an intermediate structure and rendered in React. The publishing connection lets editors change the guidance while the application owns how it appears.
Check the records behind the interfaceRecord verification and feedback to data stewards sit alongside indexing work. Improving a query cannot supply a field that is missing from the source.

The work connected policy discovery with the content and data workflows behind it. My contribution covered the search and documentation implementation, alongside collaboration with the data team on record quality.

Search quality crosses team boundaries. A confusing result might need a query change, clearer guidance, or a correction to the source record. Working through those possibilities with the people who own the data is part of building the search.