pluto is a compute cloud for exploratory financial data analysis.
- Multiple vendors, different formats, orthogonal assumptions
- Upfront investment in licensing, database setup and cleaning
- Time-stamped meta-data is non-existent
- Almost all analysis is done by practitioners who may not have a background in statistics
- Non-replicable
- Only positive backtests are shared
- Impossible to share research and solicit feedback
- High bandwidth requirement – piping different data to different endpoints
- Usage is not predictable – should a dataset be hot/warm/tepid/cold
- Data needs to be metered
- Prevent malicious code
- Prevent data leakage
- Compute needs to be metered
Who are we building for? Academic? Trader? Investor? Student?
Hard Problems + Business Model = Product Design
pluto is primarily built for the academic in you.
- Data leak-proof
- Metered database load
- Metered compute
- Facilitate collaboration
- Familiar interface
- Documented datasets
Jupyterlab is setup on the cloud where users can login with their github account, start a python or R notebook and get started. The homepage, pluto.studio, has recepies and links to working notebooks. If you run into issues, either raise an issue here or post it on slack
Read to give it a whirl? Explore on pluto.studio