What would Snowflake for an Asset Manager look like? (Part 1)
Have you ever wondered what a Snowflake Single Version of the Truth (SVOT) would look like for an Asset Manager (big bank or hedge fund)? I did and open-sourced it. I recorded this youtube playlist to demo and help you build it.
Some highlights:
- You get millisecond to 2-second response times over 2 billion rows.
- These results will be for all trades, cash used, and positions as-of any date over the past ten years. It can be used for crypto, commodities, or any other asset.
- This is 2 billion rows on one trade table. Then one window-function view to calculate the financial metrics in real-time. Notice you don’t need to materialize the results for this high-performance. This also means should you automatically stream new trades into the trade table or correct trades in the past, you don’t need to rebuild any dependent aggregations and can introduce near-real-time reporting instead of batch or over-night processing.
The biggest takeaways:
- A system like this could help big banks minimize the impact of the next big crisis like the Great Financial Crisis. This shows you how Snowflake can provide Data Governance and Data Democratization across the Front Office, Middle Office, and Back Office.
- A Cloud Data Platform for Financial Services makes every use case of an asset manager easier whether it’s for Trading, Profit and Loss (PnL), Treasury Cash Management, Risk Management, Compliance, or Regulatory Reporting.
This demo is open-sourced on Snowflake-Labs:
This youtube shows it only takes 6 minutes to build because you can use Snowflake’s instant access to compute and the instant access to the free Zepl: US Stock Market Data for Data Science dataset on the Snowflake Data Marketplace. Notice: there is no wait to spin up compute clusters and no wait to SFTP, DDL, and ETL data.
We also have a youtube deep-dive explanation of the code.
This is Part 1 of the Series:
In future stories, we will deep-dive further into use-cases and the code.
