Introduction
- Amazon QuickSight: BI tool.
- Amazon Q: a genAI-powered dashboard / report-building tool to explore data quickly.
- How QuickSight is different from Amazon Q?
- Q is a genAI capability (an separate product) and is integrated with QuickSight that allows user to interact with QS using natual language.
- In QuickSight –> Security & Permission –> Need to Enable Amazon Q.
- Need to create a Topic.
- It’s like a context / project to provide to the LLM.
- Connect to a dataset.
- Columns have to be meaningful?
- Questions:
- Do we need to provide example for Amazon Q to learn first?
- How to validate the answer provided by Q?
- E.g. in Databricks Genie, we can also see the underlying SQL query correspond to the English question and then can validate and correct it so that it can do better next time.
- There is a User Activity section where user can provide feedbacks.
- What model underneath Q?
- Use multiple models such as Nova. From tools like Amazon Bedrock.
Features
Interactive Dashboard
- AI-assisted data storytelling: tell it to generate a report based on dashboard.
- An advanced AI agent.
Ad-hoc Analysis
- Users can ask ad-hoc questions (on-demand).
Machine Learning
- Anomaly detection.
- Forecasting.
- NLQ (not realy ML)?
- Questions:
Natural Language Querying (NLQ)
Embed dashboards into applications
Misc
- Fast, in-memory engine called SPICE for efficient query performance. (Super-fast, Parallel, In-memory Calculation Engine)
- RLS
- Paginated Reporting
Pricing?
- By usage, sessions. Can charged based on roles and number of users in each role.
Data Source Connectivity
- AWS: Redshift, S3, RDS.
- On-prem: databases, SaaS application.
Compare to other tools
- v.s. Microsoft’s PBI, Databricks’ Genie: Each integrate better with their own cloud ecosystem.
- QS: UI is not as user-friendly, graph/chart library is less extensive than some rivals.
- Can’t extract underlying SQL like Genie.
- Databricks’ Genie:
- Limited on dashboard.
- PBI: