Research On Spatial Distribution Of Jobs-housing And Commuting Behaviours Based On Urban Big Data Analyzing

A tremendous amount of human mobility data has been collected by GPS enabled vehicles,automated fare collection systems and mobile phones of passengers in urban transportation systems and internet services,which can help in understanding the spatial distribution of jobs-housing status and urban commuting behaviours.The techniques of time-series analyzing are widely used in the area of speech and handwriting recognition,while it can also be applied in spatial-temporal data analyze to detect the travel and jobshousing patterns in urban areas.This study aims to find travel demand patterns and travel regularities behind bike sharing system user transaction dataset and georeferenced social network check-in dataset.This study(1)Developed a model to discretize transportation user transaction data and mobile network check-in data.Grid-based method under Lambert equal-area projection was used to discretize the spatial attribute of the dataset,while equal distance break was used to discretize the temporal attribute.This model converted the geo-referenced points with corresponding timestamps into time-series data based on grid cells.Travel demands of each grid cell were generated by this method.(2)Time-series analyze techniques were applied to analyze the discretized bike sharing system datasets.Dynamic Time Warping(DTW)algorithm was used to measure the distances of time-series,while Density-Based Spatial Clustering of Applications with Noise(DBSCAN)was used to cluster the travel demand sequences from the grid travel demand series.Five typical travel demand patterns were detect from the bike sharing system user transaction dataset,including balanced and unbalanced cases,among which the unbalanced grid cells need specific operation to avoid the uncontrolled accumulation of bikes.Furthermore,this study developed the method to evaluate the regularity of the transit users' activities.Travel frequency,departure time,travel distances,travel start and end postions and corresponding coefficients of variation of these parameters were used to describe users' regularity of travel activities.(3)The data processing and analyzing model was applied to Weibo check-in dataset,to validate the effetiveness of the model in urban data mining.The mean contribution of this study is to propose the spatial-temopral data processing model to convert user-generated and geo-referenced transaction records or check-in data into sptial distibuted travel demand sequences and activity heatmaps,allowing time-series data analyze technique to be used to uncover the regularity and typical patterns among the sequences.