10 Essential Questions To Ask For A Seamless Tableau Project
Before starting your Tableau project, you must know the key answers and insights you are trying to obtain from the data, the roles and responsibilities of users involved, and how you are going to interact with your data. Decision making in the right order can save a lot of time later by preventing myriad setbacks stemming from poor planning or communication.
These 10 questions cover the main areas you need to consider when planning your project. See the viz on Tableau Public for a list of specific questions you should ask at each stage.
1 What is the role of the Dashboard Builder?
Some users are only required to make minor edits to a dashboard or deliver specific charts. Others act as developers and designers that carry out analysis and make decisions about how to interpret and deliver insights. As roles can evolve over time, make sure to keep up to date with the level of involvement you are expected to have regarding offering solutions and driving the design of the dashboard. If not, you may make incorrect assumptions about your responsibility, miss out on important tasks, or waste effort on things that don’t concern you.
2 How do I access Data?
Ensure you have the required permissions to access data in a timely manner, or a clean line of communication with those who do. How your dashboard will be connected to data can determine some of your Tableau functionality. If you find yourself working with sample data or temporary extracts, you may find yourself losing time, or losing functionality due to switching to a data connection that does not support how you’ve been building in Tableau.
3 How is the data structured?
Know the size and shape of your data so you can plan what types of charts you can build. You may need to create joins or relationships to bring in supplementary data or use pivots or data densification to build specific charts. Be clear on the data structure before you plan and commit to a project timeline otherwise you may find that you cannot build certain chart types, or the promised chart type or analysis might require a level of data restructuring that either needs extra time to complete or slows dashboard performance.
4 What time-period does the dashboard cover?
Knowing if/how your dashboard is looking at time will determine how you build your calculations and charts. You may need parameters rather than filters or need to use LODs instead of table calculations in some places. Changing time periods further down the production line could mean you might have to rebuild complex calculations if the need for views, time filters or parameters changes.
5 Who are the dashboard users?
Is this an external client or a senior stakeholder? Will this dashboard be edited by others or open to the wider business at multiple levels? Dashboard design varies widely according to its intended use, so be aware of the requirements of the end users so you can set the appropriate level of storytelling, permissions and interactivity. You want to avoid having to rebuild dashboards to expand, condense or siphon off information, or split reports into different categories for different users.
6 What is the Data Story?
What are the key answers or insights the stakeholders need from this dashboard? Being able to interpret the desired outcome will help you to better understand and interpret specific requests. If you are not focused on the outcomes of interacting with your dashboard, or you take chart requests too literally, you may find yourself building charts that are not fit for purpose and which don’t deliver the needed insights or answers.
7 What specific content does this Dashboard need?
Do you have a clear breakdown of which areas of your dashboard serve what purpose? Are there specific KPIs, goals or targets that need to be highlighted or measured against? Being aware of these things helps you to choose the correct charts and calculations. If you begin without identifying key figures you need to extract, you may create or use the wrong dataset, miss certain data fields or promise features or reporting that proves impossible with the data you have.
8 Who drives design decisions?
Are there established design standards you should be following or are you expected to make these decisions yourself? Be clear at the outset on colours, formatting and branding so you can set up your dashboard correctly, and so you know when to adapt or push back on design feedback from other stakeholders. Being clear on the design parameters at the outset makes sure you don’t allow the project to drift leftfield due to ‘too many cooks’. It’s also a more efficient way to design as it’s possible to find the size of the project has ballooned by the time you come to manually edit colour, tooltips and unit/text fields in all your charts and views.
9 How is the project being managed?
Are there documents set up to record the brief and track feedback and development? Are all stakeholders clear of expected timelines for each stage of the project from brief to handover? Who is responsible when timelines change? Being clear on these things helps you set realistic timescales and make sure users are clear of their role and how their input affects the overall project. Poor communication or project management will mean you may miss deadlines or deliver sub-par work as crucial milestones and opportunities to salvage or pivot the project are missed.
10 How are you onboarding users to the Dashboard?
Will the dashboard be packaged and sent, shared via live connections on the server, or embedded in other locations? Will the end user have everything they need such as project documentation, definitions, Tableau workbooks or videos in order to understand and if necessary, take ownership of maintaining this dashboard after you? If you have not adequately prepared the end user to interact with your dashboards, you could deliver a product to a client or colleague that can’t be used or maintained, and thus find yourself repeatedly called back to explain, repair or amend the dashboard.
Click through the viz on Tableau Public for specific questions to ask, or to download a questions checklist.