Clinical Dashboard - Isla Health

You know that feeling when you take an idea from a scribble on paper to a fully functional tool, and you sit back thinking, I actually made this happen!?

That’s exactly how I felt about the clinical dashboard project I worked on. This dashboard was built for Isla Heathcare, A UK Health tech company that sells Clincal softwares to NHS hospitals to automate repetitive tasks. Here's a snapshot of the dataset and the link to keep you interested in the rest of the article:

Link to the dashboard

The Challenge: Data Meets Healthcare

It all started with an assignment for one of the companies I was interviewing for. Alongside the interview prep and research, I was given a task: to support a clinical team managing a post-surgical pathway. This wasn’t just any task, it was a glimpse into how data could genuinely transform patient care.

Here's the setup :

Imagine this: a patient, freshly discharged from surgery, is at home recovering. They’re on a post-surgical pathway where every step is designed to monitor their healing while minimizing complications. Here’s how it works:

1. Step 1: The patient receives a text message at scheduled intervals (e.g., day 7, 14, or 21) asking them to upload an image of their surgical wound along with a questionnaire about their symptoms.

2. Step 2: Once the patient responds, their submission creates a record in the Patient Entries Dataset. Every image or questionnaire they send is tracked individually.

3. Step 3: A clinician reviews the patient’s submission and responds with advice—whether it’s reassurance or a suggestion to visit the GP. This creates an entry in the Audit Actions Dataset, which logs the clinician’s response.

   
   

The beauty of this system? It helps catch potential infections early, reducing complications and preventing hospital readmissions. But there was a challenge: the clinical teams needed a way to track their response times, evaluate their performance, and spot trends to plan their capacity better. That's where this project came in.

The Approach: Making Data Work

Step 1: Understand the data. I had two datasets to work with:

• Patient Entries: Every patient’s touchpoint with the system.

• Audit Actions: How and when clinicians responded.

But wait—there were quirks. Missing data, invalid dates (hello, February 29th on a non-leap year!), and ghost entries. It was like untangling a ball of yarn while wearing mittens. But I powered through, cleaning and transforming the data into something usable. LINK TO DATASET

Step 2: Build the dashboard. Using Python and Dash, I turned raw data into something beautiful. It wasn’t just functional—it was interactive, intuitive, and easy to navigate. Think dynamic filters, real-time updates, and visualizations that made trends jump off the screen.

The Solution: From Data Chaos to Dashboard Magic

I built the dashboard using Python and Dash, and here’s how it turned out:

Features That Made It Shine✨

1. Median Response Time:This card displayed the median response time for selected organizations and teams. Why median? It’s less sensitive to outliers and gives a true picture of performance.

2. "Not Responded" Tracker:Identified cases where patients didn’t get a response. This static card showed communication gaps that needed urgent attention.

3. Patient Count:Displayed the total number of unique patients for any selected team or organization. Simple, yet insightful.

4. Response Time Analysis:A scatter plot that visualized response times for every patient. Hovering over points revealed details like patient ID and the exact response time. Fun fact: I found a case where it took two months to respond—definitely an outlier worth investigating!

5. Weighted Average Response Times by Organization:Showed how different organizations performed by factoring in the volume and complexity of cases. One glance, and you knew who was acing it (and who needed help).

6. Average Response Time per Team:Broke down performance by team, highlighting disparities in response efficiency.

7. Trends Tab:A line chart that tracked response time trends over months, helping teams visualize improvements—or spot where things were slipping.

8. Heatmap of Team Responses:Visualized the number of responses each team gave. Darker cells? More responses. Lighter ones? Time to dig deeper.

The Process: Making It Work

The users can:

• Filter by organization and team.

• Drill into specific trends over time.

• Access it online via Render, thanks to its seamless deployment

The Impact: Data in Action

When the dashboard went live, it became a tool that clinics could actually use. It wasn’t just about numbers, it was about people:

• Clinicians could identify where response times were slipping.

• Teams could compare performance and allocate resources more effectively.

• Most importantly, it helped patients by ensuring faster, more efficient care.

  
  

What I Learned

This project wasn’t just about data, it was about turning numbers into stories that people could act on. It taught me that even the most complex problems can be broken down into simple, actionable steps.

And honestly? Seeing something you built make a tangible difference in the world, it’s a feeling like no other.

Final Thoughts

If you’ve ever wondered how data can directly impact people’s lives, this project is my answer. From cleaning messy datasets to crafting a dashboard that clinicians actually loved, this journey reinforced my belief in the power of data-driven solutions.

The Tools: My Trusty Sidekicks

• Python: The backbone of the entire project.

• Dash: For building the sleek, interactive dashboard.

• Render: To deploy it online for real-time access.