TOWARDS THE DEVELOPMENT OF A MONITORING SYSTEM FOR PLANNING POLICY Residential Land Uses Case study of Brisbane, Melbourne, Chicago and London Presented to CUPUM 12 July 2017 by Claire Daniel Urban Planning/Data Science
ABOUT ME
THE PROJECT How can planning can take advantage of relatively recent technology for the storage, processing and display of large amounts of data (big data) in an automated and ongoing fashion for the purposes of monitoring the development of the city? Prototype focused on residential land-uses only
MONITORING IN THEORY Evaluation occurs at three stages: Before the plan Evaluation of alternative policy options and future scenarios. During the plan Ongoing monitoring After the plan Looking back at overall impacts
EVALUATION IN PRACTICE When it comes to evaluation most effort in system building is put towards predictions and projections, evaluating the future. Urban Sim SLEUTH Model Spatial Interaction Models London Land Release Model Key assumption: policy is going to work the way it is intended, that the rules will be followed
MONITORING IN PRACTICE Monitoring and evaluation into the actual effects of past planning policy is comparatively largely neglected. Where monitoring is undertaken it tends to be reported as citywide statistics without much detail
MONITORING IN PRACTICE Reasons include lack of resources, established methods and indicators, and political and organisational will. Difficulties are also found untangling the complexity of planning objectives, defining success, reconciling measurements, managing stakeholder interests and allowing enough time for policy effects to show.
CASE STUDY CITIES Chicago London Brisbane Melbourne Planning systems vary from regulatory to discretionary
CONCEPTUAL STRUCTURE OF MONITORING SYSTEM Component 1 Base Case Establish what exists Component 2 - Change Establish what has changed Component 3 Metrics How well does change meet policy objectives? Land Use Data Development Data Infrastructure Location Data Component 4 Visualisation and Communication How to present monitoring information and measurements effectively?
PRACTICAL STRUCTURE OF MONITORING SYSTEM Raw data on file Images Server R Scripts 1. Data cleaning and formatting R Scripts 2. Units of analysis and visualization methods 3. Calculation of metrics HTML, JavaScript Libraries, Google Maps API Interactive browserbased visualization MySQL to R interface with RODBC & RMySQL packages MySQL Calculated metrics MySQL Node.js express API
DEVELOPMENT DATA Establishing LAND USE what - DATA has changed QUALITY AND PROCESSING Brisbane London Chicago Melbourne Brisbane City Council Land Use Activity Dataset Ordnance Survey Address Base Plus Chicago Metropolitan Agency for Planning Land Use Inventory Melbourne City Census of Land Use and Employment Data Processing for Chicago Land use 2010 Land Parcels 2010 Census blocks 2010 Modelled land use Use Codes
DEVELOPMENT DEVELOPMENT DATA DATA Establishing Establishing what has what changed has changed DEVELOPMENT DATA - DATA QUALITY AND PROCESSING London (n=34,035) Brisbane Brisbane City Council Building Completions Certificates London Greater London Authority London Development Database Chicago City of Chicago Building Permits Melbourne City of Melbourne Development Activity Monitor Melbourne City Census of Land Use and Employment count of developments Chicago (n=11,390) Brisbane (n=15,036) What requires approval? Has it been built?
VISUALISATION CHLOROPLETH MAP Preserving and making sense of spatial patterns Dwellings developed since 2010 by ward
VISUALISATION POINT MAP Preserving and making sense of spatial patterns Location of completed schemes since 2010
VISUALISATION CLUSTERS Preserving and making sense of spatial patterns Number of Dwellings by Cluster
VISUALISATION KERNEL DENSITY MAP Preserving and making sense of spatial patterns Aggregate of kernels 1 1 1 Number of dwellings 800m Bandwidth 1 1.5 2 1.5 1 1 2 3 2 1 1 1.5 2 1.5 1 1 1 1 a b c d e
VISUALISATION KERNEL DENSITY MAP Preserving and making sense of spatial patterns
DENSITY PROFILE RELATIVE CHANGE Existing Land Use 2011 (Base Case) North Melbourne Carlton N 1km 2km N 1km 2km Percent Change (2011-2015) North Melbourne Carlton N 1km 2km
METRICS DWELLING MIX Attached and detached dwellings N 10km 20km
METRICS DWELLING MIX Number of Bedrooms N 10km 20km
METRICS PROXIMITY TO SERVICES Melbourne, proximity to parks N 1km 2km N 1km 2km Number of parks in a buffer distance Distance along road network
METRICS PROXIMITY TO SERVICES Chicago, proximity to public transport (GTFS data) Number of public transport services per stop from 06:00 to 09:00 on a Wednesday, City of Chicago Distance along road network to a public transport stop serviced on average at least every 10 minutes or more during morning peak hour (new development sites).
WEB VISULISATION
CONCLUSIONS AND FINAL THOUGHTS Hundreds of ways to program these kind of metrics Prototype monitoring system a demonstration of some of the things that can be done to get a spatially detailed picture showing which areas are relatively successful in achieving planning goals. Data attributes often reflect the administrative procedures required to process an application and are not necessarily the best for ongoing monitoring purposes. Data entry is often done by people who will not see the value of the outputs. This does not in itself define success what level or measure is considered acceptable is defined politically Achieves monitoring in an ongoing fashion, updating automatically when new data becomes available Cuts out a lot of labour intensive work otherwise making the task impossible. More testing to be done regarding usability The prototype shows that it is technically feasible, however, existing manual and digital administrative systems would need to change to make it easy in practice.
THANKYOU claire.f.daniel@gmail.com @ClaireCities https://clairedaniel.github.io/citiesdataplanning/index.html