Map a laboratory or an institution

Goal: start from the name of a laboratory, a department or a whole institution, take the people who publish there, collect their texts and build their map.

You need: cartolex installed, an internet connection (the people and their texts come from OpenAlex, an open index of scholarly works), and the name of the institution, or its ROR or OpenAlex identifier. A free OpenAlex key helps for large institutions (step 6).

It takes: a few minutes for a laboratory; for a large university, the reading of its works can take hours (see the end of this page).

Steps

1. Create the project

On the Projects screen, press Create a project. Under Start from, choose Institutions. Give the project a Name, the field’s title (what the map covers, in a few words) and a short Description (it is the context an AI reads, if you use one later), tick the Languages of the texts, and give a Folder: a new or empty folder, by its full path. Press Create the project.

You should see the People screen, on its Organisations tab, still empty, with the box People of institutions below the table.

2. Find the institution

In People of institutions, type the institution’s name (or paste its OpenAlex id, its ROR id or its link) and press Find.

The list shows every institution that bears the name, with its type, its country, its number of works, its ROR id and what it is part of: names are often shared, so choose those you mean. Tick one or several, then press Read their people.

The box People of institutions: an institution found by its name, ticked, and the authors proposed below it.

3. Read their people

A dialog asks for the First year and Last year of the works to read (empty: the project’s years) and how many Works there, at least a person needs to be proposed (two by default). Press Continue.

Before anything is sent, cartolex shows What leaves the computer: which service receives what (here, the institution’s identifiers), about how many requests, and what never leaves. Tick « I have read what leaves the computer » and press Start. (You can choose not to see this notice again for this kind of collection; Settings › Privacy shows it in full again.)

The reading runs in the background: the Activity button at the top right shows its progress, and you can keep working. cartolex reads the institution, every unit below it (labs, departments), and every work signed there in the years.

You should see the proposal: « institution: N authors with 2 works or more (M works read) », and a table of the people proposed, with their works, their years there, their units, and whether they are in the project already. A person who left keeps only the years they were there.

4. Take the people

Search the list by name, ORCID or record if you want only some of them, tick them and press Take the N selected, or press Take everyone… (N).

The dialog Take everyone proposed: the role of the people taken and the level of each type of institution.

The dialog asks:

  • their Role: Mapped (on the map, the usual choice), Context or Projected (What cartolex does, in its own words);

  • The level of each type of institution: the index’s types (education, facility…) placed on the project’s levels, so that a lab sits inside its university on the map.

When two records of the index look like one person (the same ORCID, the same name, or names that agree at the same unit), the clear pairs are taken as one person and the others are listed for you to tick. Records you know to be one person can also be selected in the list and taken with Take as one person. Press Take all.

You should see the people on the People tab, mapped, with their identity Confirmed: they were found through their records, so there is no identity to check. The Organisations tab now lists the institution and its units, with their people.

5. Harvest their texts

Press Collect, then Harvest texts. The dialog offers the years, Search HAL too (an open archive, by name) and Fill the missing abstracts from the text providers. Press What leaves the computer, read the notice (author identifiers and DOIs go to OpenAlex), tick « I have read what leaves the computer » and press Start.

You should see, when the harvest ends, « N people, M texts received » in the Activity, with Go to the build. The Texts tab lists the texts; Coverage says who has enough of them (Good: three texts with an abstract or more).

6. Build and look

Press Go to the build (or Build… on the Overview), choose the AI help (keep No AI the first time) and start the build, as in Your first map, with the demo project. On the Map, tick Organisations under Show and choose the level: each lab sits at the average place of its people. Click one: the panel lists its people, and Open in Organisations leads back to its sheet.

If something is not right

  • Too many institutions bear the name: give its ROR or OpenAlex id instead (on ror.org or openalex.org), or tick only the one whose country and type fit.

  • A large institution: its works are read 100 at a time, one request per 100 works. Above 100,000 works, cartolex stops after the first page and asks for your confirmation, with the requests, the time and the days of OpenAlex’s daily budget it needs. Narrow the years, choose units below the institution, save a free OpenAlex key in Settings › Data sources, or read a downloaded copy of OpenAlex (Very large projects: a national campaign). A reading you stop (or that meets a failing service) is paused: Resume in the Activity goes on from where it was.

  • People you do not want: on the People tab, select them and Set the role to Excluded, or to Context to keep their texts for the keywords only.

  • Two organisations that are one: « Organisations that may be one » on the Organisations tab lists them; the same ROR or OpenAlex id is merged in one step.

  • Someone without texts: their sheet (Enter on their row) gives the first blocking cause, and offers Retry, Add documents or Exclude.

More: Collecting people and texts (every way in, in detail) and Privacy and personal data.