The Framework for Data Artisans¶
Frameworks have long been the friend of the web developers, increasing productivity by easing processes of getting started, building, and deploying applications. Loosely modeled after Laravel, Carme is designed to help data scientists to create and deploy data applications.
Carme attempts to take the pain out projects by facilitating common tasks relevant to the majority of analytics teams, such as:
- Use Jupyter, Jupyterlab, and Jupyterhub setup in a container-based environment.
- Simplified version control for data and models for increased reproducibility.
- Directed acyclic graph (DAG) creation, monitoring, and deployment for data pipelines.
- Setup of cluster and GPU infrastructure for scaling analyses.
- Starter notebooks for best-of-class deep learning analyses.
- Dash and Bokeh data application deployment.
Carme can improve workflows for individuals, teams, as well as data science classrooms.