Your first project in five minutes

What five minutes gives you

Five minutes is enough to create a project, run one search and see an extracted value next to the sentence or table cell it came from. It is not enough to produce a dataset; the aim is to see how the work goes before you put a whole literature search through it.

Use a narrow topic you know well, so you can judge by eye whether the papers are being read the way you would read them.

The six steps

Each step is one page in the project menu, and each leads to the next.

• Create the project with a name and one sentence about its question.
• Write the study configuration: criteria in plain sentences, the databases to search, and target units.
• Run one narrow search.
• Screen about twenty papers, confirming or changing the suggested labels.
• Find full text for two or three included papers, and upload one PDF yourself.
• Run extraction on one paper and open a candidate value. You should see the value, its unit, its conditions and the exact place it was read from.

What you have learned

You now know whether your criteria are specific enough for someone else to apply, whether papers in your field are read cleanly or arrive as scans, and whether the properties you need are in the vocabulary.

Nothing you created is a verified value yet. A candidate becomes verified only after a reviewer approves it.