CHANGE LOG

[Terry]

November 2019 v1.2.1

• Updated to latest version of plotly library
• Enhanced procedure for rent series generation
• Fixed bug in bedroom level forecast procedure
• First version of bedroom level data in ranking table
• Overall and bedroom time-series LGA ranking tables combined

[Terry]

December 2019 v1.2.2

• Introduced suburb level store reports
• Bedroom filter introduced to suburb ranking tables

[Terry]

January 2020 v1.3

• Improved navigation on LGA pages
• Digital Store fully launched
• New site theme to improve usability and performance

[Terry]

February 2020 v1.3.1

• Started new v2 baseline code branch (alpha)
• Styling updates to ranking tables
• Free Tier Subscriptions introduced

[Terry]

March 2020 v1.3.2

• Bug fixes in forecast LGA procedure for when limited sales data is encountered
• Bug fixes due to geopandas upgrade
• Homepage layout updates

[Terry]

April 2020 v1.3.3

• Responsive changes to ranking table HTML
• Improved JSON pre-processing in ranking tables
• Introduced automated cycle detection logic
• Added beta version of Market Cycle metric

[Terry]

May 2020 v1.3.4

• Improved data clean procedure for listings
• Improved general SQL queries for better performance
• Cosmetic improvements to ranking table
• Yearly subscriptions introduced
• Updated Pricing Tables so that all features are clearly explained
• Bug fix in automated cycle detection logic

[Terry]

June 2020 v1.4

• Ranking Table JSON structure updated to improve performance
• Launched Growth Rate Cycle metric (replaced Beta Market Cycle metric)
• GRC column is now an optional toggle
• Added “Buy Now Pay Later” option to report checkout on the store
• Beta version of Tableau Reports available

[Terry]

July 2020 v1.4.1

• Bug fix where some Perth LGAs did not appear on heatmap
• Added number of rows toggle to ranking tables
• Added invoicing feature to the account dashboard
• Data for ~1,500 previously excluded suburbs added to ranking tables, charts, heatmaps
• Tableau Reports out of Beta
• Added confidence filter to Tableau Reports
• ~30 outlier localities removed from ranking tables after improved data clean procedure

[Terry]

August 2020 v1.4.2

• Tweak: Added confidence level to LGA and Suburb Forecast Charts
• Tweak: Added confidence level & colour coding to LGA and Suburb GRC Charts
• Tweak: Back-end data aggregation for Townhouse segment in preparation for next release
• Bug Fix: Incorrect confidence assignment to Devonport, TAS LGA
• UI Fix: Color coding for Median Price Forecast column on some store reports

[Terry]

September 2020 v1.4.3

• Tweak: LGA (Council Area) Confidence Levels are now calculated based on the aggregated confidence of the suburbs within the LGA boundaries
• Tweak: Changed Confidence palette for high, medium, low colour-coding to be more colour-blind friendly
• Tweak: New historical data source incorporated into the Tasmania, NT datasets
• Tweak: Minor cosmetic fixes to store reports

[Terry]

October 2020 v1.5.0

• Tweak: Suburbs with no rent or sales data are now shown on LGA (Council) pages to facilitate site search
• Tweak: Suburbs with no rental data added to downloadable reports. Rent and yield are shown as 0 values for these suburbs.
• Tweak: Removed links to LGA (Council) pages that are yet to be created on the platform.
• Tweak: Improved attribution of property sales to correct suburb by relying on GPS coordinates as opposed to suburb stated in the listing.
• Bug fix: 7 suburb forecast charts plotted median prices incorrectly.
• Bug fix: Some suburb rent forecast charts were not showing 4 and 5 bedroom series.
• Regression fix: Unit Ranking Tables had duplicate entries for suburbs with no data.
• Regression fix: Assigned 0% growth rate to low confidence outlier suburbs on heatmaps.
• Regression fix: Some suburb rent forecast charts were missing forecast figures.

[Terry]

November 2020 v1.5.1

• UI Fix: Suburbs not sorted in alphabetical order in drop-downs
• UI Tweak: New style and navigational changes on the site
• Tweak: ML model v2.0 in Alpha testing
• Tweak: Back-end changes in preparation for upcoming PSMA November 2020 LGA to suburb correspondence migration

[Terry]

December 2020 v1.5.2

• Site changes: Combined capital city and state store reports into one report with selectable options
• Data: Additional historical sales introduced into QLD dataset
• Tweak: ML model v2.0 in Beta testing

[Terry]

January 2021 v1.6.0

• • New Feature: Days on Market (DoM), Discounting, Days on Rental Market (DoRM), Vacancy Rates, Vacancies added to store reports
  • Bug Fix: Median price column was duplicated on LGA (Council Area) store reports
  • UI Tweak: Better presentation of confidence legend on GRC charts
  • Heatmaps: LGA (Council Area) and suburb shapes updated based on latest Geoscape (PSMA) Data Release
  • Data: LGA (Council Area) to suburb correspondence updated based on latest Geoscape (PSMA) Data Release
  • Data: LGA (Council Area) median values and forecasts are now calculated by averaging values for suburbs within the referent LGA. Suburbs with Very Low Confidence and/or poor past model performance are excluded.
  • Data: Approximately 3% of past property sales have been reassigned to new suburbs due to updated shape boundaries
  • Tweak: ML model 2.0 deployed in production environment. The new model has the following improvements:
    • Better handling of missing data resulting in additional time-series per number of bedrooms at suburb and LGA levels
    • Use of LGA timeseries as a regressor for suburb model train and test has been re-developed achieving a 25-30% improvement in confidence for suburb forecasts

[Author]

Posts