Skip to main content
Complementary Framework

Model catalog

Explore and test disease models

Every model here was published by a research team and checked by the framework. Open one to see what it needs, how accurate it is, and to try it with your own numbers — no account or coding needed.

Loading models…

    What the badges mean: Ready and Ready to test models can be tried with named values such as cases or rainfall. Raw input only models are not fully described yet, so they only accept a raw list of numbers and are mainly useful to the researchers who built them.

    Use the API

    Everything on these pages uses the same open API. Reading models and running predictions or forecasts needs no key. Each model page shows the exact request for the values you entered. Full reference: interactive API docs · OpenAPI specification.

    cURL

    # List published models
curl -s https://comp-api.ocular-project.com/models

# What a model needs (JSON Schema) and a ready-made example
curl -s "https://comp-api.ocular-project.com/models/<model_name>/schema?op=forecast"
curl -s "https://comp-api.ocular-project.com/models/<model_name>/example?op=forecast" \
  | python3 -c "import json,sys; print(json.dumps(json.load(sys.stdin)['requests'][0]))" \
  > request.json

# Forecast: send the example as-is, then change its values
curl -s -X POST "https://comp-api.ocular-project.com/forecast/<model_name>" \
  -H "Content-Type: application/json" \
  -d @request.json

    Python

    import requests

API_BASE = "https://comp-api.ocular-project.com/"
name = "<model_name>"

example = requests.get(f"{API_BASE}models/{name}/example",
                       params={"op": "forecast"}).json()["requests"][0]
r = requests.post(f"{API_BASE}forecast/{name}", json=example)
print(r.status_code, r.json())

    Contribute a model

    Researchers upload models through the admin portal. A bundle contains the trained model and a short description (model.json) of its inputs, so the framework can validate it and generate the test form you see on each model page.

    Want to contribute? Contact the team to request access.

    Datasets

    There are no datasets to download from this website at the moment. The models here were trained on aggregated, de-identified district data, such as monthly case counts from routine health facility reports and environmental data like rainfall and land surface temperature.

    If you need data for research or planning, contact the team. Tell us who you are, which data you need and what you will use it for, and we will get back to you.

    Please do not send patient-level information with your request.