In this episode, we speak with Tarush Aggarwal. Tarush is the founder of 5xdata, where he helps companies build a strong data foundation with self-service BI to enable the business. Prior to starting 5xData he was one of the first data engineers on the analytics team Salesforce and helped scale the data team WeWork from 5 to 100+.
The number one mistake organizations make within their data ecosystem:
Organizations try and focus on insights and gaining value from the data prematurely. Do not rush to the insights layer. Build a foundational layer. Create a self servicer layer. This will prevent bottlenecks in the future and will allow the data team to focus on moving the needle forward.
Guiding principles when designing modern data architectures?
- Data should be stored centrally (i.e. data warehouse, data lake).
- Create a data model on top of the raw data to answer 80% of your business questions.
- When organizations are first building out a data platform, the first item they should focus on is building out a self-service BI tool.
Data Warehouse vs Data Lake:
With the advancement of data warehouse’s, the ability to separate out compute and storage is a game-changer from a cost perspective. While there are still many use cases where data lakes make sense, it’s may not be the defacto anymore.