Getting started in FoodProt AI

A project is one study

Everything in FoodProt AI belongs to a project, and a project holds one study: its scope, searches, screening decisions, extracted values and final dataset. Keeping one study per project means a dataset can always be traced back to the criteria that produced it.

Projects are private. Only the people invited to a project can see it, and the role you hold in one project says nothing about any other.

Start with the study configuration

Before you search, fill in the study configuration (Scope & criteria in the project menu): the protein sources in scope, the properties you are collecting, the inclusion and exclusion criteria, and the target unit for each property.

Every later stage follows it, so clear criteria give a clear dataset.

Then work through the stages in order

The project menu lists the stages in the order the work is done. Each stage uses what the one before it produced.

• Search and import papers, then remove duplicates before anyone screens.
• Screen on title and abstract, then score the included papers by how likely they are to contain data.
• Get full text by legal routes, and upload the PDFs that cannot be fetched automatically.
• Check that each PDF was read correctly, then extract values with their evidence.
• Clear the quality checks and review every value.
• Export the verified dataset.

Two decisions are always made by a person

A person decides which papers are in scope, and a person approves each extracted value before it joins the verified dataset. Everything in between can be run again; these two decisions are recorded with a name and a time.

To see every stage filled in without running a search, load the demo project from the dashboard. It is clearly labelled and contains example (synthetic) data, not real literature.