As a data analyst, I’ve always been driven by the thrill of transforming raw data into actionable insights. But when I was tasked with creating a comprehensive Health and Safety (H&S) dashboard in Power BI, I knew I had the opportunity to create something truly impactful. This project wasn’t just about building a dashboard – it was about creating a tool that could potentially save lives and improve workplace safety. In this article, I’ll take you through my step-by-step process for designing and building an effective H&S dashboard, from defining objectives to deployment and maintenance.

A. Dashboard Design Process

1. Define Objectives and Requirements

• Stakeholder Engagement: Meet with safety officers, HR, management, and other stakeholders to understand their needs.
• Key Metrics and KPIs: Identify key health and safety metrics, such as:
  • Incident and accident rates
  • Lost Time Injury Frequency Rate (LTIFR)
  • Near-miss incidents
  • Safety compliance rates
  • Days without incidents
  • Training completion rates
  • Hazard reports
• User Roles and Permissions: Determine who needs access to the dashboard (e.g., safety officers, managers, employees) and define access levels

2. Design Planning

• Dashboard Layout: Plan the layout considering different sections for incidents, compliance, training, and trends.
• Data Visualization Choices: Choose appropriate visualizations:
  • Line charts for trends over time (e.g., incident rates over months)
  • Bar charts for comparing departments or locations
  • Pie charts for categorical data (e.g., types of incidents)
  • Heat maps for identifying high-risk areas
  • Gauges for compliance rates
• User Experience (UX): Ensure the dashboard is easy to navigate and intuitive, with clear labels and instructions.

3. Data Preparation in PowerBI

• Data Import: Identify and connect to various data sources (databases, Excel files, cloud services, etc.).
• Data Modeling: Create relationships between datasets (e.g., incidents, training records, employee data).

• Calculated Columns and Measures: Define necessary calculated fields, such as:
  • Incident rate = (Number of incidents / Total workforce) * 1000
  • LTIFR = (Number of lost time injuries / Total hours worked) * 1,000,000

4. Building the Dashboard

• Create Visualizations:
  • Incident Trends: Use line charts to show the trend of incidents over time.
  • Incident Types: Use pie charts or bar charts to show the breakdown of incident types (e.g., slips, falls, machinery accidents).
  • Department Comparison: Use bar charts to compare incident rates across different departments or locations.
  • Compliance Rates: Use gauges or bar charts to show safety training compliance rates.
  • Near-Miss Reporting: Use a heat map to show areas with high near-miss reporting.
• Arrange Layout:
  • Top Section: Key metrics (e.g., total incidents, LTIFR, days without incidents).
  • Middle Section: Detailed visualizations (incident trends, types, departmental comparisons).
  • Bottom Section: Compliance rates, near-miss reports, training completion.
• Filters and Slicers: Add filters for date ranges, departments, incident types, etc., to enable users to drill down into specific data.
• Interactivity: Enable cross-filtering between visualizations for deeper data exploration.

5. Testing and Validation

• Accuracy Check: Ensure all calculations and data representations are accurate.
• Performance Testing: Check the dashboard’s performance, especially with large datasets.
• User Testing: Conduct user testing sessions with actual stakeholders to gather feedback on usability and functionality.

6. Iteration and Refinement

• Feedback Implementation: Incorporate user feedback to refine visualizations and improve user experience.
• Aesthetic Enhancements: Adjust colors, fonts, and layout for better visual appeal and readability.

7. Deployment

• Publishing: Publish the dashboard to Power BI service, ensuring it’s accessible to all intended users.
• Access Control: Set up role-based access control, so users only see data relevant to their role.

8. Training & Documentation

• User Training: Provide training sessions for end-users to ensure they can navigate and utilize the dashboard effectively.
• Documentation: Create user guides and documentation outlining how to use the dashboard, interpret the data, and handle common issues.

9. Monitoring & Maintenance

• Regular Updates: Schedule regular updates for the dashboard data to ensure it remains current.
• Performance Monitoring: Continuously monitor the dashboard’s performance and make optimizations as needed.
• User Feedback: Maintain an open feedback loop with users to gather ongoing feedback and make iterative improvements.

B. Relationships and Cardinality Guides

S.No	From Table.Column	To Table.Column	Relationship Description
1	FactIncident.EmployeeID	DimEmployee.EmployeeID	Many incidents can be reported by one employee
2	FactIncident.DepartmentID	DimDepartment.DepartmentID	Many incidents can occur in one department
3	FactIncident.IncidentTypeID	DimIncidentType.IncidentTypeID	Many incidents can belong to one incident type
4	FactIncident.IncidentDateID	DimDate.DateID	Many incidents can occur on one date
5	DimTraining.EmployeeID	DimEmployee.EmployeeID	Many training records can belong to one employee
6	DimTraining.TrainingDateID	DimDate.DateID	Many training sessions can occur on one date
7	DimTraining.TrainingTypeID	DimTrainingType.TrainingTypeID	Many training sessions can belong to one training type
8	DimHazardReport.EmployeeID	DimEmployee.EmployeeID	Many hazard reports can be submitted by one employee
9	DimHazardReport.ReportDateID	DimDate.DateID	Many hazard reports can be submitted on one date

C. Other Important Factors to consider

1. Relevance to Stakeholders: Ensure the dashboard provides actionable insights tailored to the needs of safety officers, managers, and employees.
2. Real-Time Data: If possible, integrate real-time data to provide the most up-to-date information. e.g. Incremental refresh every 1 hours.
3. Performance Optimization: Optimize the data model and queries to ensure the dashboard loads quickly. This can include techniques like query folding, aggregations, and using DirectQuery where appropriate. e.g. Using a single SQL Connection String/ Indexing in ETL Process.
4. Clarity and Simplicity: Use clear and simple visualizations to avoid overwhelming users with too much information.
5. Consistency: Maintain a consistent look and feel throughout the dashboard for ease of use.
   • Use consistent color schemes, fonts, and design elements across the dashboard. Separate color codes for separate Modules.
6. Highlighting Critical Information: Use visual cues (e.g., color coding) to highlight critical information such as high-risk areas or non-compliance.
7. Accessibility: Use online color simulator tools. Colors to Avoid:
   • Red-Green Combination: Many people with color blindness struggle to differentiate between red and green. Avoid using these colors as the sole means of conveying information or distinguishing elements.
   • Subtle Color Differences: Avoid relying solely on subtle differences in color to convey meaning, as individuals with color blindness may not be able to perceive these distinctions.
8. Documentation and Training
   • Documentation: Provide documentation or a help guide that explains how to use the dashboard and interpret the data. e.g. create data dictionary,
   • Training: Offer training sessions for users to ensure they can effectively use the dashboard and leverage its full capabilities.
9. Compliance and Security
   • Data Security: Ensure that sensitive data is protected and that the dashboard complies with relevant data protection regulations (e.g., Australian Privacy Act).
   • Access Control: Set up appropriate access controls to ensure that only authorized users can view or interact with the dashboard.

Conclusion

Looking back on my journey, I realized that creating a successful Health and Safety dashboard wasn’t just about technical skills or data analysis – it was about understanding the needs of the people who would be using it. By putting myself in their shoes and focusing on simplicity, clarity, and performance, I was able to create a tool that would make a real difference in the lives of those who matter most. This experience taught me that data analysis isn’t just about numbers and charts – it’s about people, and the impact we can have on their lives. I hope that my story will inspire you to approach your own projects with the same passion and purpose.