Visual Analysis on High Dimensional Spatio-Temporal Real Estate Data

The analysis of the real estate market is very challenging as the data is high dimensional and has complex spatial and temporal patterns.To enable end users to create their own visualizations and gain insight into the real estate market,a novel HTML5-based visual analytics approach which integrates four interactive visualizations is proposed.The method includes a map view to show the geographical information of houses using location-based clustering,a stacked graph view to show the evolution of house sale over time with combinations of various layouts and orderings,a pixel-bar view to visualize multiple attributes of houses using sale and price-based clustering,and a treemap view to present the hierarchal structure of the data with combinations of various nodes layouts and orderings.Rich user interactions are further proposed to enhance the flexibility and analytical ability of the entire system.We have applied our method to the real property market data and obtained some interesting findings.The feedback from the end users of our system is very positive.