03 / Academic project / Urban data / Walkability / 2025

Sydney green space.

A spatial data prototype for exploring parks and investigating walking connections.

GreenSydney map with park filters and location markers
01 Park exploration. The map presents locations and park attributes in an interface for comparing places.

The design question

Can park data become useful spatial information?

I brought park records into a Python and Flask application so locations, attributes and contextual questions could be explored together.

Project outputs
  • Park map with filters and information panels
  • Analysis dashboard and data-processing workflow
  • Park-specific contextual question endpoint

Select an image to enlarge it and read its caption.

Context
Academic project / Spatial data prototype
My contribution
Data processing, mapping, Flask application & contextual AI integration
Tools
Python / pandas / Flask / JavaScript / Google Places / OpenAI integration

01 / Interface & analysis

Move from an overview to specific questions.

Filters and information panels let users focus on park attributes such as location, size and facilities. The analysis page presents charts and recommendations, while the question endpoint can incorporate the details of a selected park.

This demonstrates how a spatial dataset can be made easier to interrogate. It also shows why the origin of each indicator matters: pedestrian estimates in the prototype are synthetic, and walkability scores are derived rather than measured assessments.

Analysis dashboard with pedestrian activity charts and park recommendations
02 Analysis interface. Charts demonstrate the prototype using synthetic pedestrian estimates; they are not a live foot-traffic feed.
Proposed green-space data architecture connecting sources processing and interfaces
03 Proposed system architecture. The diagram describes the broader concept, including components beyond the implemented prototype.

02 / Data & system design

Build a path from records to an explorable map.

Python processing prepares park records for use by the application. A companion Google Places integration handles search, detail requests, enrichment and merging with local CSV records before exporting JSON.

The system diagram connects data acquisition with mapping, analysis and contextual questions. It is useful as an explanation of the intended architecture, while the screenshots and source demonstrate the prototype’s implemented components.

The data-enrichment script starts in test-data mode. The AI endpoint needs configured API access; this portfolio presents its design and implementation rather than a live assistant.

03 / Communication & limitations

Make the data’s status visible.

A credible urban dashboard needs to explain which information is sourced, estimated or simulated.

Project overview with navigation to park mapping and analysis
04 Project overview. Mapping and analysis are presented as complementary parts of the application.
Data-quality panel identifying source data and synthetic indicators
05 Data quality. The interface distinguishes park records from synthetic pedestrian data and derived walkability indicators.

The outcome demonstrates how urban data can be processed and communicated through a spatial interface. Before informing decisions, a future version would need verified pedestrian observations, consistent access measures and testing with the people using the dashboard.

Image detail