John Hawkins's Projects
Search Algorithm for Protein Structures
Data preparation tool for preparing machine learning datasets to build academic literature filtering tools
An experiment to build and populate a Feature Store Group using a Sagemaker Pipeline
A small collection of custom kernels for running Sagemaker Notebooks an Training Jobs
Exploratoration and Experiments on designing custom Workbenches in AWS Sagemaker
Automated Bayesian Forecasting
An End to End Customer Churn Prediction solution using AWS services.
A presentation on the nature of Bayesian Reasoning
Simple
An experimental variation of the GBM algorithm
Simple scripts for generating synthetic data
A simple library that uses the DataRobot Python API to generate 2 way partial dependency plots
Application for re-ordering DataRobot Leaderboard by Aggregated Error Metrics
An example application of how to build a calibration model to improve probability estimates for a machine learning model
Examples of clustering and visualising the prediction explanations returned by DataRobot
Application for doing partial dependency on DataRobot models that contain features that have co-linear relationships
Meta modelling tool for taking a very large dataset and gradually building a series of DataRobot models to maximise the use of the data
Demonstrating how to run a multi-target model factory and automatically create a containerised dashboard
Example of using DataRobot to build a quantile regression model
Demonstration of using DataRobot AutoML to build survival functions using censored data about expected time to unrecoverable events.
Examples of how to build custom workflows in DataRobot
A set of reusable methods to transform datasets for machine learning.
A library and command line application to generate a table of summary stats on a columnar dataset. Produces results as either markdown or latex
Web Application for building simple enzyme efficiency models from FASTA sequences
Data sets containing generic calendar events useful for time series models in Australia Business
Bits and pieces of code for running numerical and statistical experiments