The Brief
Rental hosts face challenges in optimizing pricing and predicting occupancies due to volatile markets and shifting seasonal demands. This project was developed during my postgraduate studies at NBCC to analyze public rental datasets and turn raw tabular files into structured dashboards answering host pricing questions.
Approach
I structured this build into a modular data pipeline:
- Data Extraction: Loaded sprawling public CSV data representing city listings, host metrics, and seasonal occupancy records.
- Cleansing & Normalization: Wrote a Python script using Pandas to identify missing values, handle empty descriptions, remove price-currency outliers, and clean geographic coordinates.
- Modeling: Loaded the normalized tables into Power BI, establishing logical relations and writing DAX formulas to calculate rolling averages and host ranks.
The Stack
What I Built
A multi-view Power BI dashboard mapping listings, host performance, and occupancy trends across seasonal fluctuations. Dynamic filters allow users to filter down by neighborhood, property type, or rating grade to see matching metrics in real time.
Outcome
Presented the final model to instructors and peers. Surfaced key insights, such as occupancy drop-off rates on weekdays versus weekends and listing price correlations to specific amenities, demonstrating strong BI modeling and data handling skills.