Cognite’s Industrial Data Science Library

This is Cognite’s collection of data science algorithms and models. Its objective is twofold. First, empower domain experts to conduct exploratory work, root cause analysis, and analyze data without the requirement to code by driving the industrial data science and analytics engine behind Cognite Charts. Second, curate a collection industrial relevant data science algorithms to be used by any data scientist. For more information, check Charts documentation page.

Installation

To install the core part of the library, where dependencies are numpy, scipy` and pandas, run:

pip install indsl

InDSL also includes extra functionality that that requires additonal dependencies. The extra functionality includes:

  • Numba, for pre-compiled functions and faster execution

  • Plot, for plotting functionality that depends on matplotlib

  • Modeling, for functionality that depends on csaps and kneed.

  • Stats, for statistical functionality that depends on statsmodels

  • Scikit, for machine learning functionality that depends on scikit-image and scikit-learn

  • Fluids, for fluid dynamics calculations and simulations.

To install extra functionality, run:

pip install indsl[numba,plot,modeling,stats,scikit,fluids]

To install all the extra functionality, run:

pip install indsl[all]

Toolboxes

Examples

Developers

Indices and tables