What is our primary use case?
We tried to do a proof of concept with QuickSight. Essentially, we were evaluating it for visualization to create charts and graphs to explain the reports that were being generated behind the scenes. Instead of having to go through hundreds of thousands of lines of tables, we just wanted to get a high-level summary and then figure out the areas in which we needed to go deeper. We used its latest version.
How has it helped my organization?
It didn't. We weren't able to fully implement it due to security concerns.
What needs improvement?
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.
For how long have I used the solution?
We tried it for a three month period.
What do I think about the scalability of the solution?
We were selecting it for scalability, but we didn't get that far. Our clients are usually large businesses. They are multi-million dollar government agencies.
How are customer service and technical support?
I didn't contact them, but I heard that there were some issues with response time, or the answers that they were getting were not quite where they needed to be.
Which solution did I use previously and why did I switch?
I initially built the solution by using data tables and Databricks. We were initially building out prototypes just to find out what the client wanted and to determine the BI tool to proceed with. At the end of the day, if we had a better developer who was able to easily wire up the Databricks data tables, high charts, and interactivity, it probably would have been the solution. The problem is that when you have a group of data sets, how are users supposed to know which data set they need to incorporate into the build to get the results that they need.
There is some way that QuickSight could do that and say, "Here are all your data sets that you have, and here's the content inside of it, or at least the high-level entities that match." It would be much easier to be able to say, "Hey, I have an entire Data Lake that is full of information, and I want this kind of answer." It should then be able to direct you to the data set that you need to incorporate to get those answers.
How was the initial setup?
I didn't do its initial installation and setup, but I know that it was not straightforward. It could have also been because of the quality of developer we had. It wasn't just a straight plug and play, especially when you have a multiple-tier architecture. It would be great if there was like a line from the CDN that you can just install inside of either NPM or install it on the backend to be able to populate with API on the front end, but we didn't really have the best developers working on that. I know that they had some problems with the implementation, whether because of datasets that were populating it or something else. It wasn't straightforward. Based on their struggles and the fact that they couldn't turn it around in a proof of concept in a day, I would say that it was not easy to implement.
If you took Excel and you knew pivot tables, you could go in and easily create a pivot table out of the Excel data within. Even if you don't understand it, it is pretty straightforward, but QuickSight was not as straightforward.
What's my experience with pricing, setup cost, and licensing?
If it is on an iframe on a separate server and we're trying to employ something that is serverless technology, we have to pay each time they went and tapped the service. So, every call to the server had a cost associated, which can make it quite expensive depending on how much you use. I've dealt with smaller hosts with unlimited bandwidth and unlimited space for 20 hours a month. The prices of AWS, Azure are astronomical. It is kind of frustrating, but it is what it is. It is the industry standard.
What other advice do I have?
I would advise others to have quality developers and ensure that the security patches had been in place and be mindful of that moving forward so that if there are any security issues, you can call it out before the client calls it out.
I would rate Amazon QuickSight a five out of ten.
Which deployment model are you using for this solution?