Introduction

This quick-start guide will help you get started using the VizUrban Portal. For best experiences, the portal is only supported for major modern browser such as latest Google Chrome, Internet Explorer 11, Mozilla Firefox and Opera. To check the compatibility of your browser with the portal, please refer to this link.

About VizUrban

A Volumetric Visualization tool for Multivariate Heterogeneous Urban Data

The big data from cities, if managed and visualized well, can be used by authorities and urban designers and planners to inform urban renewal and development projects. The volume and variety of big data present challenges in presenting analysis of collected data for human to cognitively perceive valuable information, particularly in the dynamic and complex urban contexts. This project investigates techniques for aggregating, filtering, analysing, and visualizing multivariate federated data sources. A web-based tool for visualizing different variables in up to four dimensions is implemented on two major cities: Adelaide and Brisbane, as demonstrator sites.

Access and Navigation

VizUrban Portal Log In

In the landing page of http://www.vizurban.com, a user credential is needed in order to access the VizUrban portal. Once you have acquired your user credential, please input username and password then press Enter or click the "Login" button to authenticate.

If you wish to get an access to the portal, please contact the AURIN team.

After successfully authenticating the account, you will be redirected to visualization of polygons on default geographical boundary on region level of Statistical Area(SA) 1 for Adelaide.

Changing maps

The default map provider can be changed by clicking the map icon on the top right corner of VizUrban portal. In this case, we are going to change the map provider to Bing Map Roads.

Side Navigation Bar

Notice that the portal is facilitated with side navigation bar to assist in your data visualization.

This side navigation bar can be hidden by toggling the small blue button "<<" on the top right side from the side bar, this button would also trigger the side bar to be visible later on.

Since we are now visualizing polygons on SA1 region level, we are going to change this to Suburb region level. Open the polygon feature in the side navigation bar by clicking on "Polygons" title header.

A section of polygon rendering tool would appear in response to the action we just performed. Select "Suburb" in dropdown options and click "Get Polygons" button.

After the map finish reloading the polygons, the shapes are changed to suburb from SA1 polygons visualization as shown above.

For the visualization of polygons, we can further zoom in or out by scrolling the mouse on the map overlay to get clearer view for geographical boundary rendered. Get Polygons action can then be triggered again manually to get expanded visualization of polygons on current camera view of the map. The rendering of polygon boundary is set based on elevation of current window view of the portal.

The orientation of horizontal view for the map can also be changed by middle click action on mouse, hold and drag the camera view to get different viewing angle as desired. This function can also be triggered through keyboard key "Ctrl" + mouse click and drag actions on the map view as shown below.


Additionally, the opacity level of rendered polygons can be changed through the opacity slider.

In this case, we are going to change opacity level to 0.7 (70%) to get clearer and solid polygons view which result is shown in the following screenshot.


A quick navigation tool is provided in side navigation bar in order to make easy redirection onto specific area of interest.

Current area of interest provided in VizUrban visualization tool consists of Adelaide and Brisbane (Logan). Any click action performed on these controls would change the camera view to target area location and requests for polygon rendering on selected region level in the Polygon tab.

Log Out

Use the following feature to log out from the portal through "logout" button on top right corner of side navigation bar

Metric Datasets Visualization

In VizUrban visualization tool, several datasets are provided for states of South Australia and Queensland. Brief tabular information of each dataset is presented below:

Metric Dataset Unit Colour Height Data Coverage
Population Number of people in region Supported Supported South Australia (2011)
Population Density Population/sqm Supported Supported South Australia (2011)
Transport Frequency Total public transport services/week Supported Supported South Australia and Queensland
(until 31 January 2015)
Transport Coverage Total public transport stops in region/week Supported Supported South Australia and Queensland
(until 31 January 2015)
Transport Supply Transport Frequency/Transport Coverage Supported Supported South Australia and Queensland
(until 31 January 2015)
Land Usage Percentage of land use type Supported Supported South Australia (2011)
Housing Price Median housing price ($AUD) Supported Supported South Australia and Queensland
(sales in 2014)
TransitCare Total transport origin or destination per month Supported Supported Queensland
(1 month data from November 2014)
Transport Delay Average public transport delay in seconds Supported Not Supported South Australia and Queensland
(15 December 2014 - 31 January 2015)

Data Sources

Metric Dataset Providers Notes
Polygons for SA1, SA2, SA3, SA4 and Suburb Australian Bureau of Statistics Data is publicly accessible
Population AURIN
Population Density AURIN and Australian Bureau of Statistics Aggregation from Population and Polygons datasets
Land Usage AURIN
Housing Price AURIN
Transport Frequency Adelaide Metro and Translink Data is publicly accessible
Transport Coverage Adelaide Metro and Translink Data is publicly accessible
Transport Supply Adelaide Metro and Translink Data is publicly accessible
Transport Delay Adelaide Metro and Translink Data is publicly accessible
TransitCare TransitCare

Visualization

There are two types of metric visualization in VizUrban portal identified as Colour Metric Visualization and Height Metric Visualization. Each metric visualization type is able to perform the dataset visual projection onto rendered polygons of the map. These are accessible through the navigation side bar of VizUrban portal.



Case Study 1: Transport Frequency in Commercial Area

This case study will spefically target analysis of commercial land use of region with public transport frequency.

To inspect the relationship between public transport frequency and commercial land use of polygons, we are going to use combination of color and height visualization.

Select Land Usage option for the dataset dropdown control in Colour Metric tab. Alternative dropdown selection for land use types would appear after changing option from 'No Metric' to 'Land Usage' for the dataset.

Continue to change the selection from Agriculture to Commercial for land use type metric then press on "Display" button.



The projection of color metric shows heatmap visualization on the suburb regions. The small red region in below figure shows that the particular suburb has highest percentage of commercial land use type.




Notice that infobox in above figure appeares right after the tool finishes rendering color heatmap visualization. It gives basic statistical information on given dataset of commercial land usage such as mean, median, standard deviation and value range.

Zoom closer to the red area and select the red suburb area by clicking on the polygon to highlight the area. Area selection infobox will appear immediately on top right corner of portal map revealing that the red suburb "Green Fields" has commercial land usage for approximately 38.4%. Additional information of all available land usage types of selected suburb would be presented.



Upon displaying the color metric visualization color range legend will be shown on the right hand side of the map. This legend will allow clarification of value range with its color hue in heatmap projection on rendered area polygons. As shown below, the projection of color for commercial land usage dataset has 38.40 as the highest value mapped to red color indicating the hottest area. Blue hue color in heatmap visualization will indicate regions having low density data value (less significant) whereas white would indicate regions with no data provided in selected dataset.



Next step is visualizing the transport visualization dataset. The current objective for this particular analysis is to see which commercial suburb has high transport frequency.

To achieve the objective, height metric visualization is used for projecting transport frequency dataset.
Open the Height Metric tab and change the selection from 'No Metric' to 'Transport Frequency' for dataset field and click "Display" button.



When the visualization tool finishes rendering, projection of transport frequency dataset will be shown as height of each suburb polygon as described in the following figure.



*Notice the green height visualization statistic infobox appears on the screen giving brief information for mean, median, standard deviation and value range of transport frequency dataset within the same restricted geographic boundary as color metric projection.

Visualizing transport frequency data as height 3D polygon projection shows clutters on the screen due to high range value within given dataset, therefore we are going to normalize the height using z-index rescaling function through height metric tab.

Open height metric tab on the VizUrban side bar and tick the "normalize" checkbox then click on "Display" button.





As the result, the view will show different projection after normalisation process of height for the suburb polygons. What changed from previous process was only the display of polygon heights, the raw data provided from dataset is not affected.

Afterwards, select the polygon that has the highest height. This will again reveal "Adelaide" suburb in selected area infobox having commercial land usage percentage as 6.9% and transport frequency of 346,619 services/week.



Given with the situation that Adelaide Suburb having highest transport frequency, questions are asked in relation to "Green Fields" Suburb with high density value for commercial land usage percentage.

To assist data visual analytics in this case study, several features are offered by VizUrban visualization tool to filter out some polygons in which condition we are not interested with.

Each metric visualization tab is given value range filter slider and toolbox referred as "Statistical tools" upon projection of the dataset. Below figure shows the filtering options for land usage projection for color metric tab.



For our analysis referring to Green Fields Suburb, a filter should be applied to quickly see the suburb with high percentage of commercial land usage. Change start range of slider filter from 0 to approximately 20.20. The polygons which have percentage value of commercial land usage will then disappear from the visualization.





Here we can find the red polygon area where Green Fields suburb is. If we do inspection by clicking on it, the selected area infobox reveals that this specific suburb has only 1626 total services per week.

*Note that the selected area can be cleared by selecting non polygon area or double clicking the selected suburb, different actions can also be triggered using statistical tools to quickly set the value range slider for filtering the metric dataset.

Considering that Green Fields suburb has low transport frequency despite having highest commercial land use, it is suggestive that infrastructure needs to be improved in that specific suburb by adding more public transport services. However, there might be a reason why public transport services might not be crucially needed due to other factors. Next section will continue the analysis from this case study to assess the relation between population density and its transport frequency.

Case Study 2: Transport Frequency with low population density

Continuing from previous case study of commercial land usage's transport frequency, we are going to perform in depth inspection if green fields should have more public transport services to be added.

First, change the selection of color metric dataset from commercial land usage to population density and click on "Display" button. All color metric filter we previously performed would be reset and adjust to new value range based on population density dataset.



Given with new heatmap visualization for suburb population density against normalised height of transport frequency, statistical tool will again be used to filter out regions to quickly find Green Fields suburb. In this case, change the size of standard deviation to 1 and click on button "< Standard deviation". This will result to visible polygons are only rendered whose population density is below calculated 1 standard deviation in metric dataset. As shown in below figure, Green Fields suburb only has approximate 4735 people per sqm as value of its population density.



Due to the fact that Green Fields having low population density, the argument in case study 1 that needs of public transport services might be impractical in this case.

To assist decision making towards the needs of adding public transport services with taking both commercial land usage type and population density into consideration, a different approach in visual analysis is necessary.

Drag color value range filter endpoint to the maximum value to reset the color projection of population density dataset then go to heigh metric visualization tab in side bar, tick reciprocal checkbox and finally click on "Display" button.

The visualization would then give different height projection by taking normalised value and performs inversion function on visual data. Therefore, what we see in below figure reveals high height projection suburbs will have low public transport frequency.



Next, untick "include no data" option in height metric tab in left navigation side bar. This will then hide suburbs in which transport frequency data is not available.



From the underlying visual analysis, we might conclude that Kingston Park suburb needs special attention for public transportation services. Despite having relatively high population density of approximately 1,421,432 people per sqm, the public transport services that pass through that suburb is noted as 75 total services/week.

Furthermore, in order to correlate the population density and transport frequency with commercial land usage, a filtering feature for land usage types is provided in the left navigation side bar.

Set the land use types filtering condition to "any" then deselect other land use types in dropdown options except for commercial. To process these conditions, click on the filter button.

Unfortunately Kingston Park suburb is hidden due to the fact that the suburb does not have any commercial land use in place.



*Note that there are 3 possible conditions of land use types filter.

  1. "all" condition will only show polygons having all selected land use types specified in filter.
  2. "any" condition will only show polygons having any of selected land use types specified in filter.
  3. "only" condition will only show polygons having only selected land use types specified in filter.
Additionally, this filter feature is only feasible to be applied onto Adelaide areas. Brisbane is not applicable due to no available dataset for land use types.

Case Study 3: Transport delay analysis

Transport delay analysis is a relatively special dataset only available to color metric visualization. Temporal analysis can be performed on the top of color heatmap projection.

To start with, open color metric dataset option and set it to "Transport Delay". Interface for showing date range and delay aggregation options would be shown just below the dataset selection.



For this case study, we are going to select date range betweeen 1 January 2015 to 31 January 2015 with weekly aggregation option selected.



*Notice that the statistical and filtering for color metric won't be available for transport delay analysis. This is intentional to reduce dimension complexity of temporal analysis. In addition, a time slider of "Variance Metric" will appear after color metric projection of transport delay dataset.



With the time slider described in above figure, you can perform visual temporal analysis by observing changes of delay on rendered polygons as illustrated below.



Case Study 4: TransitCare Correlations

About TransitCare

TransitCare is an innovative not-for-profit organisation dedicated to providing community transport for eligible customers in the Brisbane South and Logan areas. More details can be found here

TransitCare metric dataset

As private data provider, TransitCare provided 1 month historical data to be used only for visualization demonstration of VizUrban portal.

Visualization and Correlation Options

It is important to note that the TransitCare dataset is only available for Logan (Brisbane) area, thus this can be easily navigated through "Quick navigation" tab to change the camera view location to Logan, Brisbane.



Next, change color metric dataset selection to "Transit Care". A dropdown option would appear beneath the dataset dropdown control to select which particular dataset (origin or destination) you want to visualize on color metric projection.
For this demonstration, please select the transport origin and click "Display" button to proceed.



Correlations visualization for VizUrban is apparently only available for TransitCare dataset. This could only be served via selected area of color metric visualization of TransitCare dataset.
Select one of the polygons on the map. This will show selected area infobox with additional inline interface for correlation options as illustrated below.



Based on selection of TransitCare origin selection in color metric tab, there are two possible selections of correlation provided:

  1. Correlation to Public Transport Route Frequency.
  2. Correlation to TransitCare destinations of selected area.

To show the correlation, hit on the inline "Correlate" button in selected area infobox. Projection of correlated dataset would be shown as normalised polygon heights.
Below figure describes side by side comparison of correlation options for selected area we chose earlier.



Data Export

Several datasets of rendered polygons are exportable through VizUrban portal. Notice the export button on Polygons title tab on left side navigation bar on portal is provided to export polygon shapes and datasets associated to a GeoJSON file.


Once you click on the export button, a modal box will appear to let you choose what datasets you wish to export along with the rendered polygons. Once you are satisfied with selection datasets to export, click on "Export as GeoJSON" button. The browser will begin the download for exported file.


*Note that TransitCare and Transport Delay datasets are unavailable to export.

Visual Pattern Recognition: Heatmap Visualization

As extra feature for heatmap color metric visualization, color redistribution is offered in color metric tab in order to achieve cognitive pattern recognition for visual analytics.


There are 5 distribution mode can be selected for visual color redistribution:

  1. Normal
    Raw data is accordingly mapped to color hue proportionally (regular linear color map).
  2. Low
    Logarithmic scale from lowest value, visualization potentially displays low value polygons as significant areas.
  3. Mean
    Absolute logarithmic scale from middle value (up and down), visualization potentially displays polygons having dataset value closer to mean as significant areas.
  4. High
    Logarithmic scale from highest value, visualization potentially displays high value polygons as significant areas.
  5. Extreme Skew
    This distribution mode is similar to High mode. However, Extreme Skew mode may filter down some of significant polygons and re-identify them as less significant areas.


Disclaimers

This project was made possible with funding from the Australian Urban Research Infrastructure Network (AURIN)

Acknowledgements

Private dataset providers



Agreement between VizUrban and TransitCare

Information provided by TransitCare was restricted to 1 month historical data which the data is not exportable through VizUrban portal. Provided data is only used for visualization purpose within VizUrban portal.

Project Team Overview

Chief: Dr. Flora Salim, School of Computer Science and IT, RMIT University.

Team members:

  • Irwan Fario Subastian
  • Jason Linayage
  • Jonathan Liono
  • Samuel Elsmore
  • Soewanto Anggono Ang
  • Zachariah Duncan Lee