I don't like the fact that we had to implement it using an iframe. To implement it on a website, you have to use an iframe. It would be better if we could call up an API to populate content on the website versus having your alignment iframe. This is because when it is on a separate server, you have to deal with cores. 

I also couldn't customize the look and feel to meet the environment in which we were trying to install it. Its initial setup was also not straightforward.

We weren't able to fully implement it due to security concerns. One part of our company is currently implementing Databricks because of security issues with AWS QuickSight.

It's not a Tableau. It's not a TIBCO Spotfire. They could do further integration with some data governance and data cataloging. You can't tell a data story as well in QuickSight as you can on others, especially in TIBCO Spotfire. If they integrated it more with some data cataloging and had some guided analytics where it will walk you through everything, that would be ideal. Most clients don't expect it to do that. That said, once again, it isn't a Tableau. The results are as good as Tableau, and yet, it doesn't, from an engineering perspective, tell the data story or allow a user to be able to access a data catalog that lets you drive how your analytics need to go. I don't if they're working on that, however, that would be a lovely enhancement.

Amazon QuickSight has minimal features. The feature set can be improved to allow you to create good production-ready dashboards. This does not happen with QuickSite, it needs to mature to be used in production.

With Amazon QuickSight or any other BI tool, they are limited in what you can do because they're drag and drop tools. For unique features we need or what we want to build, we cannot build using these tools. We use JavaScript for that. We write our own code and build our own solutions.

I need the capability to call machine learning models in Python. For example, while I'm building a dashboard if I am displaying a chart, but the value calculation should be a machine learning model, which is running somewhere else, such as on Amazon. I need that capability because these tools give good outputs, such as calculated fields. However, today the outputs are not straightforward. It's not only some additional or multiplication, but you also need a machine learning output to come, and then you want to show it. These tools cannot do that on the fly. In simple, you can say machine learning on the fly is not currently provided.

The visualizations that it has got are not as robust as some of the other competitors in the market. We would like to have more visuals. 

The modeling side of it is not as robust as Power BI.

It should support integration with Python or R.

Its support should be improved. There isn't much support for this solution.

QuickSight is gradually adding new features, but it is not as mature as some of the other products on the market such as Tableau or Power BI.

The cost is something that could be improved, it should at least be competitive at least for a year until they get a good hold on the market penetration.

As I have not yet explored the solution to its fullest, It may already have this available, but what I would like to see is better, drag and drop based integration of the data sources into the Amazon QuickSight dashboard, so that you can automatically integrate it with different types of data cleaning tools.

There could be an end-to-end pipeline for data cleaning as well as presenting it using different visualizations.

It is a simple tool with limited features. Its visualization set is very limited, and it also has limited functionality. An intelligence tool should not be only for creating reports. Currently, we have to do all computational and mathematical operations outside. We should be able to do such operations in an intelligence tool like this.

As an intelligence tool, it should support dynamic refresh. QuickSight currently supports the refresh at a minimum of one hour, and it is not suitable for dynamic dashboards that require frequent refreshes. 

There is room for improvement in terms of the number of visualizations and dashboards that are available. 

There should be better connectors for different data sets.