v2.0 Use Knoema instead of deprecated Zepl Share; Added Python Faker for synthetic trader creation; Rewrote Snowsight Tiles
Removed: company_profile, stock_latest, finservam_datascience_wh, snowflake_sample_data, share_now, middleware
v1.0 Use Zepl Marketplace share
Big banks and Asset managers have spent millions of dollars to accurately give a Single Version of Truth (SVOT) in real-time. What would such a system look like on Snowflake?
Snowflake has high performance at low TCO due to its out-of-the-box operation, near-zero maintenance, and low learning curve
SVOT makes trading, risk management, regulatory reporting, and Financial Services big data use cases significantly easier
Unlimited Compute and Concurrency enable quick data-driven decisions
Use the Snowflake Data Marketplace to instantly get stock history so the business doesn't have to wait for IT.
Populate only the trade table and use window functions to generate cash, positions, and Profit-and-Loss (PnL) so that you can have real-time updates.
Use SnowSight - Snowflake's complimentary User Interface (UI) - to generate dashboards that can be shared with the business.
Youtube Demo and build-from-scratch
Find a share named "Knoema Economy Data Atlas" from the Snowflake Data Marketplace and mount the database as economy_data_atlas
Run Script 10: Sets up the environment
Run Script 20: Connects to the Data Marketplace to get free stock history
Run Script 30: Populates the trade table. Creates Window Function Views for cash, positions, and PnL
Run Script 40: This is the smoke test and what the business queries
In line 27 of script 30, you can set the variable limit_trader = x.
It's default is set to 100 traders to populate in the trader table which when multplied by 40+ years of daily trades will create 3-billion-plus trades.
| limit_trader | Trades generated | Script 30 Run-time with xxlarge compute |
|---|---|---|
| 100 (default) | 3 billion | under 4 minutes |
| 200 | 6 billion | |
| 300 | 9 billion |
Data Build Tool (DBT): Open-Sourced by a hedge fund prospect. DBT gives you software engineering best practices on big data with concepts like Don't Repeat Yourself and Analytics Engineer.
Sigma Computing: Excel-like analysis over 2 billion rows powered with only Snowflake Small compute power.
Run "optional\finserv 90 reset.sql".