My approach to create Tableau Dashboard
Firstly I would like to thank you Kate Strachnyi, for giving me this opportunity to share my views and experience of creating Tableau dashboards.
As you might know data in today’s world is becoming vital and “The world is going to be data”. I think this is just the beginning of the data period.
There are many BI tools using which you can sense your data and make your business decisions more intelligently then previously, however one should be good enough to actually to use the capabilities of data analytics and business intelligence tools.
Tableau is one of the leading tools which is meant for analytical and BI reports development.
When I want to understand the pain points of my data or tell a story out of data, I prefer Tableau because of its easy to use features.
So how do I start working on my Tableau projects or what process I follow?
Well, when I use Tableau in my day job; I try to understand with the stake holders or report end users what are they trying to look out of Tableau dashboard, what is there business, what are the business requirements and few more follow up questions.
Once I gather the grounds for the project, I start to look for how are they currently utilizing the data, ex. is it an Excel report. what dimensions and measures do they want, which filter options they want for the end users.
And then prepare a document for the business requirements just for reference purpose.
Then comes the data source gathering, mostly the data happens to be stored in multiple data sources or coming from flat files.
I try to understand how the data is captured, what is the granularity of data, are all the required fields present in the data source which I require in my dashboard.
Then comes the data preparation part, where I would perform activities like cleansing, modifying the columns data type and transforming using ETL tools. I have used SSIS and Tableau Prep for these activities. Although there are some really good options available.
Once the data is pulled, I check with stake holder for data validation (Most important part of process I think) maybe sample of data.
Then follows the step of actual creating visualization into Tableau desktop.
Add those dimensions and measures to the canvas per business requirements. Creating calculated fields for those fields which aren’t readily available.
Tableau provides 24 charts by default in SHOW ME menu, however we can create some amazing custom charts in Tableau, by making some data preparation.
Once I have finalizd the view, I try to look for improving the performance of dashboard, this is important when you are dealing with huge amount of data.
If I have to mention them, using extract of data rather then using Live connection (Both have its own advantages and disadvantages) , use of context filter, try avoiding exclude option while filtering, filtering those fields that are not used (Custom SQL query at your help).
Following of some best practices likes of using same color code across all views.
I think there are 2 types dashboard :-
Depending upon my requirements I try to create the dashboards.
Last thing when you have create a dashboard and its live for use, I try to document the fields used, formulas for the calculation fields, data sourced used for future reference.
I hope you like my approach and look forward to read and hear more about it to improve my approach building the dashboards in Tableau.
By: Chaitanya Kulkarni