Ok, this one seems obvious. However, since I am writing about it, you can guess, that it’s not.
The guidance I would like to give you is to utilize the Tableau provided connectors in cases where it exists. For some of you data-savvy folks, you might have a very good understanding of ODBC connectors and in many cases, you might already have ODBC setup on your machines to connect.
We highly recommend utilizing the option that Tableau provides when you click on the name of the vendor
So, in the example of Oracle, we would want you to use the “Oracle” option and NOT THE “ODBC” option
First read this to understand my perspective on Sets, then read this to understand all sorts of cool things you can do with Set Actions.
Then, sprinkle your dashboard with some Set Actions. This is like a bit of salt or lemon on your dish, just a bit. Don’t overpower your dashboards with too many set actions or you are going to leave your consumers with a disgusted taste
Viz in tooltips got a huge applause when it was launched at the TC a few years ago, but I tell you, I am not applauding the analysts who are now wanting to use this in every dashboard they build.
There is a specific purpose to building dashboards and that is to give your consumers ability to see things at a high level and yes, also have the ability to slice and dice the data using interactivity and filters.
Viz in Tooltip is a great way to get that next level of understanding that goes with the context but is it a good way to do adhoc analysis. That is the issue when my customers are asking for putting dashboards inside of Vizzes. I am not sure about that.
Any how, here is a way I think of using this features
So, this might a decent example of using Viz in Tooltip
But, this is NOT (as it requires too much thought processing from the consumer).
This raises many eyebrows. Why does the number of data sources impact my dashboard. And, I would say, it shouldn’t. And we have made many advancements where it might not.
However, in some of the older version prior to 10, we saw a direct co-relation between the number of data sources and the initial load time it took for a dashboard to load on the Tableau Server. If you are seeing that having a higher number of data sources is causing an issue, I would advise you to create dashboards that don’t use more than 3 data sources and compare it.
In any case, if you are combining more than 3 data sources for a single dashboard, I really think you should be using your enterprise data store (EDW or similar) to get the data so you aren’t combining data in real-time with data sources that are very different in terms of their technologies.
Another thing to consider is that Tableau makes separate connection to each one of the data sources so if you have 29 data sources to a database, Tableau might be doing some of the metadata work 29 times.
This happens to fall in the last category where there are certain items that are completely out of control for business teams.
However, having additional information hasn’t even hurt anyone. So, if time and again, you just need to ask your IT counterparts if the servers has enough hardware and it keeping up with the load of the users and the usage, they may appreciate that. In fact, you might need to prepare for some funding too so you can not only ask the right question but offer to help provide $ needed to ensure your success.