Machine Learning
MediSight AI
An explainable healthcare risk-modelling workflow built around model evaluation, experiment tracking, and a clinician-facing interface.
Case study · Applied AI
Predictly
Property intelligence · Web app
A real-estate valuation dashboard that combines market context with transparent, local image-quality signals.
01 / Context
The problem
Property estimates are easy to present as precise even when inputs are incomplete. Predictly separates its estimate from image-derived signals and makes the supporting context visible.
My role: Personal project; implementation and limitations are described in the public repository.
02 / Implementation
How it works
01Capture the property context
The interface gathers location and property details to build a structured estimate request.
02Inspect image quality
Pillow-based image processing calculates visual properties such as brightness and sharpness. These are image signals, not a learned property-valuation model.
03Present a range of context
Maps, trend views, and a report help users inspect the estimate and its assumptions in one place.
Implementation highlights
- Property estimates for Lahore, Karachi, Islamabad, Rawalpindi, Faisalabad, and Multan
- Local image analysis for brightness, colour richness, contrast, and sharpness
- PDF report, historical-trend views, map exploration, and mortgage estimate
- Comparable listings and amenity adjustments are presented as contextual features, not independently verified market data
03 / Evidence
What can be verified
The points below summarize the public repository and owner-provided CV where noted. Read the source for implementation details and the evaluation setup before drawing further conclusions.
Image analysisLocal processing
Pillow-based image features; no external vision API is required for this analysis.
Coverage6 cities
Cities listed in the project README.
ReportPDF export
The app includes a property report workflow.
04 / Boundaries
Limitations and next questions
- The visual features are not a trained computer-vision valuation model.
- Comparable listings and market context may be illustrative or mocked; the estimate should not be treated as a certified appraisal.
- Property prices change over time and require verified, current market data for real decisions.
Evidence-first note: a live demo demonstrates an interface and workflow; it does not by itself validate model quality, security, or real-world impact.
Continue exploring