This is a collaborative project aimed at analyzing hospital data to uncover weaknesses in the healthcare system. The ultimate goal is to support decision-making that improves service quality, ensures better doctor and department management, and identifies areas where additional resources are needed.
- Detect weaknesses in the hospital system.
- Evaluate doctors’ and departments’ performance.
- Support better resource allocation (doctors, services, departments).
- Provide actionable insights to improve healthcare services.
- Excel → Data cleaning and preliminary analysis.
- SQL → Data extraction, combining tables, and KPI calculation.
- Power BI → Interactive dashboards and reports.
- Python (Colab) → Advanced visualizations & machine learning.
- Machine Learning → Predictive models for emergency visits & patient satisfaction.
- Some doctors with low ratings still serve a large number of patients, indicating possible workload imbalance.
- Departments such as Pediatrics and Emergency are understaffed and require more resources.
- X-ray is the most requested and most expensive procedure → should be prioritized in insurance coverage.
- Emergency visits often represent ~50% of total visits, requiring continuous staffing and service readiness.
- Costs remained stable overall, but showed anomalies (e.g., drop in October due to fewer visits).
- Hypertension is the most frequent diagnosis among patients aged 60+.
- Outpatient services are the most common service type.
- Predicting Emergency Visits → Used feature selection & multiple models to achieve the best accuracy.
- Predicting Patient Satisfaction → Tried regression & classification approaches. Classification performed better after converting satisfaction into categories.
⚠️ Note: The dataset is not real hospital data (dummy/simulated).
- Open the Colab notebook: Google Colab Link.
- Install required libraries (Pandas, Matplotlib, Scikit-learn).
- Run the notebook to reproduce visualizations and machine learning models.
Here are some examples from Power BI & Python Visualizations:
