top of page

Insurance Underwriting & Risk Performance Dashboard- Power BI

  • benlusic
  • Jun 18
  • 3 min read

Updated: Jul 24



Building a P&C Insurance Underwriting Intelligence Dashboard: A Journey in Iterative Design

When building data analytics tools, it is easy to assume the hardest part is writing the initial data model or SQL scripts. But as any analyst will tell you, the real battle is won in user experience (UX) and visual hierarchy.

Recently, I built a Property & Casualty (P&C) Insurance Underwriting Intelligence Dashboard in Power BI. It is designed to give insurance underwriters and sales leaders a bird's-eye view of policy volume, lead conversion rates, and risk distribution.

The final product looks polished, cohesive, and executive-ready—but it didn't start out that way. Here is the honest breakdown of how this project evolved, the mistakes I made along the way, and how iterative design saved the dashboard.


A Power BI data analytics dashboard titled "Source of Lead" and "Account Profitability." The top section features horizontal button slicers for filtering by lead source (Agent, Online, Referral) and risk profile (Critical Risk, Highly Profitable, Moderate Risk). The middle row contains three charts: a pie chart showing "Leads: Won Vs Lost," a donut chart tracking "Conversion Rate by Source," and a bar chart detailing "Policy Count By Credit Score" ranging from tiers 1 to 5. The bottom section displays two comprehensive data matrices—"Claims Frequency Vs Credit Rating" and "Policy Type Distribution by Age Group"—styled with alternating green and gold rows, alongside a prominent KPI card displaying a total "Premium Revenue" of 22M.

Link to the Interactive Dashboard: Power BI


Part 1: Spotting Redundancy and Streamlining Layout

My initial layout featured a classic wide-to-narrow "Time to Conversion" funnel chart right in the middle of the page. It tracked our lead pipeline: Online, Agent, and Referral.

However, looking at the top of the canvas, I already had a prominent button slicer allowing users to filter by those exact lead sources. Right next to the funnel sat a donut chart displaying Conversion Rate by Source.

[The Mistake]: Visual Redundancy. 
The funnel was taking up prime real estate just to repeat information the user could already see in two different places.

The Pivot: I realized that if a chart doesn’t introduce a completely new dimension to the data, it's just noise. I stripped out the funnel entirely and replaced it with a Policy Count by Credit Score column chart. This introduced a crucial risk-profiling angle that perfectly complemented the underlying financial data, giving underwriters an instant visual on portfolio risk exposure.


Part 2: The Missing Data Trap (Power Query Realities)

Once I added the new credit score bar chart, I noticed a jarring gap on the X-axis: Credit Tier 4 was completely missing, leaving a blank space between Tiers 3 and 5.

[The Mistake]: Hardcoded conditional logic mismatches. 
When creating the custom sorting keys in Power Query, I mapped the number "4" to the text label "Excellent". 
But when I looked closer at the source data, the actual category label typed in the rows was "Very Good". 
Because of that spelling mismatch, Power Query skipped the tier entirely and dumped those records into a catch-all bucket.

The Pivot: I jumped back into the Power Query Editor, opened the conditional column steps, and explicitly mapped out all five tiers to perfectly match the raw dataset strings (Poor, Fair, Good, Very Good, Exceptional). After fixing the conditional sorting logic and updating the visual to display the friendly text labels instead of raw numbers, the missing bar snapped back into place.


Part 3: The Formatting Matrix & Combating "Color Fatigue"

The final hurdle was the visual styling. I wanted a theme that felt professional and trustworthy, so I leaned into a corporate dark blue and green palette. But color can easily lie to your users if you aren't careful.

[The Mistakes]: 
1. Lack of Contrast: My canvas background and chart backgrounds were identical, making the entire dashboard look completely flat.
2. Accidental Color Meaning: I used the same solid "Success Green" for won leads, multiple matrix rows, and the Premium Revenue KPI card. This accidentally signaled to the viewer's brain that the total $22M revenue belonged ONLY to won/highly profitable accounts, rather than representing the global portfolio total.

The Fixes:

  • Creating Depth: I dropped the master canvas background to a soft, premium light gray and toggled on solid white backgrounds for the individual visual cards. This made the charts cleanly "pop" off the page.

  • Matrix Transparency: I stripped out the master backgrounds of the data tables and set the grid options to transparent, allowing them to sit natively on the canvas. To fix readability, I swapped the heavy cell fills for soft, subtle, alternating desaturated tints.

  • Neutralizing Global KPIs: I changed the Premium Revenue card back to a clean, neutral tone with a soft drop shadow so it stands on its own as a global metric.


Key Takeaways

Building this dashboard reminded me of a core data analytics truth: the dashboard is a story, not just a collection of charts.

Fixing alignment issues, removing duplicate visuals, and ensuring that a color means the exact same thing everywhere it appears are the details that separate a confusing report from an executive-ready tool.


Key Project Specs

  • Tool Used: Power BI Desktop

  • Core Concepts: Data modeling, Power Query conditional transformations, UX/UI dashboard layout design, color theory hierarchy.

 
 
 

Comments

bottom of page