# NannyML/nannyml

nannyml: post-deployment data science in python

Repository: https://github.com/NannyML/nannyml
Canonical: https://ross.abutalabs.com/products/nannyml
Homepage: https://www.nannyml.com/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: machine-learning, ml, mlops, performance-monitoring, data-science, monitoring, python, data-drift, model-monitoring, data-analysis, visualization, deep-learning, jupyter-notebook, machinelearning, postdeploymentdatascience, performance-estimation
Last push: 2025-07-12T08:12:42+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 31, release rhythm 16, longevity 100
- inputs: {"age_days": 1680, "days_push": 417, "days_rel": 417, "gap_med": 130, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2150, forks 190 (observed 2026-08-28T04:06:19.282020+00:00)

## What it is
NannyML is an open-source Python library for post-deployment machine learning monitoring. It estimates model performance without access to ground-truth labels, detects univariate and multivariate data drift, and links drift alerts to changes in model performance.

## Use cases
- estimate ML model performance without ground truth labels
- detect data drift in production ML models
- monitor classification and regression model degradation
- link data drift alerts to model performance changes
- detect multivariate drift in feature distributions
- visualize model performance over time in Jupyter notebooks

## When to choose
- you need to monitor deployed ML models when ground truth is delayed or unavailable
- you want a model-agnostic, Python-native monitoring library for tabular data
- you want to avoid alert fatigue from drift alerts that don't affect performance
- you prefer an open-source alternative to commercial ML monitoring platforms

## When to avoid
- you need monitoring for image, text, or video models rather than tabular data
- you want a fully managed hosted monitoring service without self-managed infrastructure
- you need real-time streaming monitoring out of the box rather than batch analysis

## Facets
- artifact type: library
- maturity: active
- function: monitoring, machine-learning, data-visualization, analytics
- domain: machine-learning, data-science, monitoring
- platform: python, cross-platform
- tags: ml-monitoring, data-drift, performance-estimation, model-monitoring, post-deployment, mlops, concept-drift

## Member repositories
- NannyML/nannyml (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:19.282020+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:50:57.357753+00:00, confidence not recorded.
  - readme: https://github.com/NannyML/nannyml (fetched 2026-08-28T04:06:19.282020+00:00, sha acc406f3427c)
  - homepage: https://www.nannyml.com/ (fetched 2026-08-29T10:30:53.657126+00:00, sha 5e7703c55def)
  - site_page: https://docs.nannyml.com/cloud (fetched 2026-08-29T10:30:53.659854+00:00, sha f5378d5b2037)
  - site_page: https://www.nannyml.com/about (fetched 2026-08-29T10:30:53.677772+00:00, sha cd6e7a7ff0eb)
  - site_page: https://docs.nannyml.com/cloud/model-monitoring/quickstart (fetched 2026-08-29T10:30:53.680079+00:00, sha 2af626ba73bc)
  - registry_pypi: https://pypi.org/pypi/nannyml/json (fetched 2026-08-29T10:30:53.682079+00:00, sha 483746c3c324)
  - site_page: https://www.nannyml.com/pricing (fetched 2026-08-29T10:30:53.661955+00:00, sha 845caf4e2ae8)
- Data as of 2026-08-30T08:39:29.467469+00:00.