Python Data Analysis and Machine Learning Services
Pubrica provides data analysis and machine learning in Python for research groups, healthcare organisations and life-science companies. Python is widely used for data processing, statistical modelling and machine learning. Our data scientists use pandas, statsmodels, scikit-learn and related libraries, and deliver documented code with validation results that a reviewer can follow.
- Documented code delivered as scripts or Jupyter notebooks with a requirements file
- Validation reported on data the model was not trained on, with calibration as well as discrimination
- Leakage checked by fitting preprocessing steps within the training data only
- Reporting to TRIPOD+AI where a clinical prediction model is developed or validated
- Data governance agreed before work starts where patient data are involved
Python scripts or Jupyter notebooks with a requirements fileIncluded
Trained model files where applicableIncluded
Validation report with performance and calibrationIncluded
Methods and results sectionsIncluded
Data processing logIncluded
Service Scope
What Our Python Analysis Service Covers
Each project is scoped individually. The components below are combined according to your study and its requirements.
Applications
Typical Applications of Python
Process
How the Project Works
Methods, deliverables and fees are confirmed in writing before work begins.
Yes. The code, the environment specification and any trained model files are delivered with the report.
Performance is reported on held-out or external data, with calibration and the behaviour at clinically relevant thresholds.
Yes, under a data processing agreement and the approvals your study requires. Where possible we work with de-identified data.
Yes. Where a method or a regulatory requirement favours another package, we recommend it and agree the choice with you.
Pubrica Academy
Further Reading in Pubrica Academy
Request a Quote for Python Data Analysis
Send your dataset description and research question. We confirm the methods, the deliverables and the fee before any work begins.