Find Collaboration Opportunities

Discover institutions you should be working with based on shared research topics

Which institutions are leaders in your fields that you haven’t yet collaborated with? This recipe identifies who publishes in your top research areas, checks how much you already co-author with them, and flags rising stars. We’ll use MIT as the example throughout.

Part 1: Topic-based discovery

Find institutions publishing heavily in your research areas that you have few co-authored works with.

Step 1: Find your top research topics

Group your institution’s works by primary topic to see where you publish most:

https://api.openalex.org/works?filter=authorships.institutions.id:I63966007&group_by=primary_topic.id&per_page=5

[
  {"key": "https://openalex.org/T10048", "key_display_name": "Particle physics theoretical and experimental studies", "count": 2753},
  {"key": "https://openalex.org/T10346", "key_display_name": "Magnetic confinement fusion research", "count": 1994},
  {"key": "https://openalex.org/T10299", "key_display_name": "Photonic and Optical Devices", "count": 1888},
  {"key": "https://openalex.org/T10039", "key_display_name": "Stellar, planetary, and galactic studies", "count": 1820},
  {"key": "https://openalex.org/T10037", "key_display_name": "Physics of Superconductivity and Magnetism", "count": 1680}
]

Step 2: Find who else publishes in those topics

For each topic, find the top institutions by output:

https://api.openalex.org/works?filter=primary_topic.id:T10048&group_by=authorships.institutions.id&per_page=10

[
  {"key": "https://openalex.org/I67311998", "key_display_name": "European Organization for Nuclear Research", "count": 14034},
  {"key": "https://openalex.org/I1294671590", "key_display_name": "Centre National de la Recherche Scientifique", "count": 7824},
  {"key": "https://openalex.org/I1314696892", "key_display_name": "Fermi National Accelerator Laboratory", "count": 6823}
]

Step 3: Check existing collaboration levels

Use the + operator to count co-authored works between two institutions. The meta.count in the response is all you need:

https://api.openalex.org/works?filter=authorships.institutions.id:I63966007+I1294671590&per_page=1

{"meta": {"count": 4970}}

MIT and CNRS share 4,970 co-authored works — a strong existing relationship.

Step 4: Spot rising institutions

Compare who’s publishing in a topic now versus five years ago. Run the same group_by query for two time windows:

# Recent (2023–2025)
https://api.openalex.org/works?filter=primary_topic.id:T10048,publication_year:2023-2025&group_by=authorships.institutions.id&per_page=10

# Historical (2018–2020)
https://api.openalex.org/works?filter=primary_topic.id:T10048,publication_year:2018-2020&group_by=authorships.institutions.id&per_page=10

Institutions that appear in the recent top 10 but not in the historical list are rising fast. For particle physics, the University of Chinese Academy of Sciences jumped into the recent top 10 with 672 works but didn’t appear in the 2018–2020 list at all.

Part 2: Find the gaps

Now compare who you do work with against who you could work with. A single call gives you all your co-authoring institutions at once:

https://api.openalex.org/works?filter=authorships.institutions.id:I63966007,publication_year:2020-2025&group_by=authorships.institutions.id&per_page=100

[
  {"key": "https://openalex.org/I136199984", "key_display_name": "Harvard University", "count": 7525},
  {"key": "https://openalex.org/I107606265", "key_display_name": "Broad Institute", "count": 3122},
  {"key": "https://openalex.org/I97018004", "key_display_name": "Stanford University", "count": 2319}
]

Institutions that rank highly in your topics (Part 1) but don’t appear in your co-authorship list — or appear with low counts — are your biggest opportunities.

Full script

This script combines both parts: finds your top topics, identifies who leads in each, pulls your co-authorship list in one call, and highlights the gaps.

import requests

BASE = "https://api.openalex.org"
MY_INST = "I63966007"  # MIT

def api(endpoint, params):
    return requests.get(f"{BASE}/{endpoint}", params=params).json()

# Get all co-authoring institutions in one call
collab_data = api("works", {
    "filter": f"authorships.institutions.id:{MY_INST},publication_year:2020-2025",
    "group_by": "authorships.institutions.id",
    "per_page": 100,
})["group_by"]

collabs = {g["key"].split("/")[-1]: g["count"] for g in collab_data}

# Get top 5 topics
topics = api("works", {
    "filter": f"authorships.institutions.id:{MY_INST}",
    "group_by": "primary_topic.id",
    "per_page": 5,
})["group_by"]

for topic in topics:
    tid = topic["key"].split("/")[-1]
    print(f"\n=== {topic['key_display_name']} ({topic['count']} works) ===")

    # Top institutions in this topic
    top_insts = api("works", {
        "filter": f"primary_topic.id:{tid}",
        "group_by": "authorships.institutions.id",
        "per_page": 15,
    })["group_by"]

    # Recent vs. historical for rising star detection
    recent = {g["key"].split("/")[-1] for g in api("works", {
        "filter": f"primary_topic.id:{tid},publication_year:2023-2025",
        "group_by": "authorships.institutions.id",
        "per_page": 15,
    })["group_by"]}

    historical = {g["key"].split("/")[-1] for g in api("works", {
        "filter": f"primary_topic.id:{tid},publication_year:2018-2020",
        "group_by": "authorships.institutions.id",
        "per_page": 15,
    })["group_by"]}

    rising = recent - historical

    for inst in top_insts:
        iid = inst["key"].split("/")[-1]
        if iid == MY_INST:
            continue

        coauthored = collabs.get(iid, 0)
        tag = " ** RISING **" if iid in rising else ""
        print(f"  {inst['key_display_name']}: "
              f"{inst['count']} topic works, {coauthored} co-authored{tag}")