The coding tool for scientific and engineering time-series data.
Your domain experts shouldn't have to spend their time writing Python code by hand. FeatureGraph handles the time-series analysis so your team can focus on interpreting the results.
I'd like an analysis built on my data
We set up FeatureGraph on a historical extract of your time-series data and answer an agreed set of questions. You receive a checked report and a record of how every result was produced.
- A report written for your domain experts
- Result tables you can use in your own tools
- A provenance record, so every result can be reproduced
- Raw data is not retained after the pilot
I'd like to build my own software
We provide the data store, agent layer, and the analysis backend built using featuregraph-core, the open-source compiler. Everything you need to integrate FeatureGraph into your workflow.
pip install featuregraph-core (MIT-license)
- Documentation for the framework
- A public demo of the question-to-query loop
- The research record and published studies