Answers generated from Webex transcript and chat interactions
Based on the transcript, I’ll format the Q&A portion in the Community Q&A format:
Q: Is it possible to add data to scenarios directly from datasets instead of from a dashboard or files? E.g., can we do scenario planning on data for which a dataset exists but no one has built a dashboard?
A: Currently, you can only plug in data from dashboard visuals and static files uploaded from your local desktop (Excel or CSV files). Direct dataset connection is not supported at this time, but it is on the roadmap.
Q: If my instance of QuickSight doesn’t have scenarios, is that a permission restriction on my user permission?
A: Scenarios is only available to QuickSight Pro users (Admin Pro, Author Pro, and Reader Pro).
Q: In which AWS regions are Q scenarios available?
A: Q scenarios are currently supported in:
- US East 1 (Virginia)
- US West 2 (Oregon)
- EU West 1 (Ireland)
- EU Central 1 (Frankfurt)
Q: Does the scenario consider both datasets or just one when answering questions?
A: It considers both data sources that are provided.
Q: What happens if I run out of thread but want to continue?
A: You can start a parallel thread since there’s no limit on the number of threads you can create. The only limit is 15 blocks per individual thread.
Q: Can we connect to other data sources except uploaded files (e.g. Athena or S3)?
A: Currently, you can only use dashboard visuals and files (Excel/CSV) uploaded from your local desktop.
Q: How do you prep the dataset (example of labeling columns and cleaning data)?
A: Within scenarios, you have three main data prep controls:
- Creating filters on the data
- Skipping columns that aren’t relevant
- Sorting the data
The actual data prep should happen before plugging data into scenarios.
Q: Is row level security (RLS) supported in scenarios?
A: Currently there is no support for row level security in scenarios.
Q: Does it help if you explain the dataset to Amazon Q?
A: Within scenarios, you don’t need to explicitly explain the dataset to Q. Scenarios has its own intelligence layer and relies on the dashboard setup to detect patterns, correlations and generate alternative solutions.
Q: Can you provide more details about the LLM/machine learning model behind the scenes? Is it capable of joining data points or creating new calculated fields?
A: The model is built on Amazon Bedrock. It does have the ability to create new calculated fields (as demonstrated by creating an “average sleep quality score” from other sleep-related parameters). Due to the proprietary nature of the model, additional technical details cannot be provided.