I am new to Tableau, I have been trying to understand it and work with it for 3 weeks now. It seemed to me, from my first impressions about it, that many different employees inside organizations use it in their work and that it takes from them an hour or so to build a dashboard. This means it is both handy and fast but in my own experience up to now it needs a lot of time reading and understanding and experimenting.... a full time work. I wonder about the process of how it is adapted in organizations.
Now I have some data I want to use Tableau with, I am reading a lot to try to understand how I will connect this data to get the desired dashboard. I have some conceptual questions whose answer will help me get the clues to make progress- finish data connections to start building dashboards- and save time.
- If I want to design a dashboard for live data, should I build it in the first place with extracts or live? What is the better or optimal way? After live connecting the the first data source, when I try to connect to the second live data source, the query request time starts to count without ending; I have no other way except shutting down the computer to solve it. I wonder how we can in principle connect hundreds of live data sources and work efficiently in worksheets at the same time?
- I have a folder with some .csv and excel files, now here two inquiries: first, what if I want to "blend" these files? They are all already in the same database, but every time I want to add a file from the same folder for blending I have to open the file as a new data source and get again all the files contained in this new data source. Is this reasonable? Cannot I blend from the same folder opening it only one time as one data source? the same data to blend a new file I get the page of opening a new data source?
Actually these files are describing the performance of some measures in time, e.g. months. So they actually share different column headers. So I want to add them to each other as a rows below each other, not next to each other, and still preserving the measures that are not available for all of them. I wonder about the clue of achieving this very simple purpose?
- What if I need to blend around the date by days, I read that this not very efficient. I basically want to compare all the details of different data sets on the same day.
What if I want to blend by months to see monthly performance of each product not the monthly average to them all together!
What if I want to blend by a sub category (all individual categories have the same segmentation. e.g. category one has a b c category 2 has the same a b c but of category 2.... etc.) again to see the details of this sub category for all categories each not one value describing the average of them all?
Thank you so much,,