We have developed open-source infrastructure–we call it Mangrove–that allows to create and maintain the dataset of open reviews, for everyone to be able to use it freely.
The different building blocks are available under Apache License 2.0 and you can inspect or contribute to the code here.
Technical standard to ensure interoperability and allow reviews to be analyzed programmatically (data format, etc.)
Open-source API and documentation (REST API with an OpenAPI specification), allowing for easy submission and retrieval of reviews
Client-side JS library allowing interacting with the Mangrove server and managing of reviewer identity
Open-source probabilistic aggregation algorithm to ensure reliability of the data
Demo web app to showcase the capabilities of the infrastructure, allowing to read and write reviews for places (points of interest), legal entities, websites, books, and more
For the Open Reviews Ecosystem to succeed and the infrastructure to be maintainable by the community, we believe the underlying technology has to ensure three main qualities:
The data format has to be useful for as many parties as possible, so that the dataset is valuable and attractive to use.
Capture as many aspects as necessary for different use cases
Open and free-of-charge access
The infrastructure needs to be easy to integrate into websites and apps for reading and writing reviews, so that the ecosystem gains integrators and the dataset grows.
Clear technical standard that ensures interoperability
Easy-to-use and well-documented APIs
Clear legal framework to provide confidence in the integration
The infrastructure needs to be easy to maintain at low cost and little overhead.
Reliable and persistent database server
Open source development to allow for redundancy and open innovation
Effective governance structures to ensure the infrastructure can evolve as the needs of the ecosystem participants change