Toolkit for data science and machine learning built by Spotfire®

This package provides several functions for Data Science Machine Learning (DSML) capabilities. While the functions can be used for any Python code they have been optimized for use in or alongside Spotfire. The functions include:

• Train pipeline-based supervised models
• Pipeline-based pre-processing and modeling for text data. Topic modeling, information retrieval, and advanced clustering analysis
• Preprocessing, smoothing, SAX encoding, lag/rolling window statistics and univariate forecasting for time series data
• Pattern exploration using matrix profiling
• Perform model explainability
• Calculate model drift
• Distribution fitting, normality testing, parameter estimation, distribution prediction, and probability prediction
• Missing data summary, comparison, removal and imputation methods for tabular data
• Geo-analytics capabilities in the form of coordinate transformations, distance analysis, shape generation, spatial joins

Installation

pip install spotfire-dsml

License

BSD-type 3-Clause License. See the file LICENSE included in the package.