A dashboard is a critical tool for understanding an organization's business performance. Decision-makers may see their company's key performance indicators (KPIs) via a single interface. The successful implementation of a dashboard is difficult and necessitates a step-by-step procedure — a methodology that takes into account all parts of the project lifecycle.
A simple dashboard development process would include the following steps:
1. Initial Requirements Gathering Phase
In this phase, the BI Developer does preliminary analysis and gains functional knowledge about the team through a series of interviews. He also gets to know his stakeholders, defines clear ways of working and gathers requirements using a Dashboard Planning Document ( or equivalent Requirements Doc). The main questions that he needs to answer are - What is the purpose of the dashboard? What are team processes? Get a list of the main metrics and KPIs. What is the target audience? Data Frequency Update and many more. Please feel free to check out my Requirements Gathering Article. Always ensure to sign-off requirements after they are gathered so both parties agree on paper exactly what should be delivered.
2. Design - Wireframes!
There is a saying - 'A picture paints a thousand words!'. As part of the second phase of the project, you should ensure that you design the dashboard using a design tool such as (Miro, draw.io or any other tools that you can find online or your organisation uses). When drawing your dashboard designs take into consideration:
The visualisation your stakeholder wants (Line chart, Pie chart, Bar or Column charts or any other).
Make sure to list all the filters that will be affecting the visualisation
Ensure that you don't add more than 10 visualisations on a tab/page/dashboard
Conduct a series of meetings or interviews to showcase your progress to your stakeholder and ensure that you are capturing all of their requirements. Once you have agreed on the designs, always ensure they are signed off!
3. Dashboard Development
Depending on your organisation (whether you are required to do data modelling or not) this phase would start differently. In this article, I will cover the approach where the BI developer has to do data modelling as well.
3.1. Model your data correctly and document it!
Start by identifying the sources of your data! Make sure that you have clearly mapped them.
Build ETL processes that will extract the data from the sources and transform it into tables hiding all GDPR and PII data and aggregating if needed then load the result into fact and dimension tables.
Always ensure documentation is provided. Simply add an ERD to show relationships between the fact and dimension tables.
Ensure that the data is correct by showing your stakeholder the result and asking him to test some of the raw data (if they are data-savvy!).
Have as many interviews/checkpoint meetings as possible to ensure that the data is correct.
3.2. Dashboard Development
Use the dashboarding tool that your organisation is using to build the solution (Power BI, Tableau, Sisense or any other BI Tool out there).
Produce a first prototype of the project which is based on your Wireframes and showcase that to your stakeholders! - What usually tends to happen is when they see something already built they start asking for additional requirements, some of which might be out of scope. Make sure those out-of-scope requirements are recorded as part of the delivery of the project but remind them that both parties have already signed off and agreed on requirements and designs. In other words, accept only requirements that you think will not massively affect the deadlines of the project.
Once you have made changes all the changes that the stakeholder wants you will have your final prototype which should be ready for testing.
4. Testing
The testing phase is an extremely vital part of every project in IT. It ensures that not only the requirements are met but we get rid of any bugs in the system. In order to fully test your project these are the recommended steps you should take:
Individual Testing of your data and dashboard - Use a dashboard and data quality checklists to test your project in a methodical manner ensuring that you record the date of the tests carried out. Simple checklists might be a good idea to test both your data and dashboard. You can find more information in the Dashboard Testing Article!
Always Peer-Review! - Always be humble when doing a project and consider that we're human and we make mistakes. Getting your colleagues to review your work and critique it is the only way to achieve excellence. Ensure that the peer-review date and name of the person have been recorded in case you need to come back to it.
UAT - Get your stakeholders involved in the testing phase of the project. They are the ones that are most familiar with the processes and data and would know if something is wrong straight away. Involving them in the Testing phase makes them feel appreciated and ensures that your project meets all their requirements. You can reflect also on the Requirements and Wireframes documents which you signed off in the initial phase of the project.
5. Deployment and Monitoring
The last phase of the project is ensuring that your project is productionised in the appropriate location (Folder of Production Reporting System). Also, ensure that the dashboard is shared with all the users that need to have access to it!
Provide documentation on your project as part of the delivery. You can find more information on the type of documentation in this article.
Once released ensure that the project is maintained and over-time additional requirements are added to expand the dashboard and bring more value to the department/team.
6. Training
A common mistake that most BI Developers do is not training your core users on the dashboard you built. Assuming that your users know how to use the BI tool and know how to drill down and slice and dice your data is a huge mistake. Always ensure that you have trained your audience on how to use the tool. Some suggestions for when you're planning your training session might be:
Have a template PowerPoint presentation where you show them how to access the system and how to request access to the report if they don't have it already!
Do a live demo and demonstrate how they can slice and dice the data and filter the charts and tables within your dashboard.
Include your stakeholder in the presentation to walk them through his user story and show them how he benefits from this dashboard. It's the perfect way to engage everyone and sell your project.
Ensure that you leave out enough time for questions from the audience
Thanks for reading this! I hope you find it helpful. Love to hear your feedback and queries if any!
_edited_edited.jpg)
Comments