Interactive Analysis of Large-Scale Spatial Vector Data via Display-driven Computing (DEMO)

To address the scale issue for analysis of large-scale spatial vector data, we present a new spatial analysis computing model, display-driven computing (DisDC), for interactive analysis of large-scale spatial vector data. As show in Fig 1, DisDC generates analysis results by directly determining the value of each pixel for display. Different from data-driven computing (DataDC) (see Fig 2) in traditional analysis methods, the computing units in DisDC are pixels rather than spatial objects. As the number of pixels in the screen range is limited and stable, DisDC has the advantage of being insensitive to data volumes.

Fig1. Spatial analysis via display-driven computing

Fig2. Spatial analysis via data-driven computing

Setting

Tab1. Datasets of Demo 1: Roads, POI and Farmland of Mainland China (10-million-scale)

Name	Type	Records	Size
China_Road	LineString	21,898,508	163,171,928 segments
China_POI	Point	20,258,450	20,258,450 points
China_Farmland	Polygon	10,520,644	133,830,561 edges

Tab2. Datasets of Demo 2: Spain Datasets from OpenStreetMap (Classification standard)

Name	Type	Records	Size
Amenity_Education	Point	6,994	6,994 Points
Amenity_Entertainment, Arts & Culture	Point	6,928	6,928 Points
Amenity_Financial	Point	13,040	13,040 Points
Amenity_Healthcare	Point	14,757	14,757 Points
Amenity_Sustenance	Point	89,282	89,282 Points
Amenity_Transportation	Point	3,114	3,114 Points
Shop	Point	82,192	82,192 Points
Religion	Point	6,219	6,219 Points
Historic	Point	14,974	14,974 Points
Leisure	Point	8,334	8,334 Points
Tourism	Point	36,462	36,462 Points
Places	Point	582,464	582,464 Points
Public_Transport	Point	45,478	45,478 Points
Railway	Linestring	97,675	309,716 Segments
Highway	Linestring	3,132,496	42,497,196 Segments
Waterway	Linestring	33,214	4,254,732 Segments
Natural_Beach	Point	148	148 Points
Natural_Cave	Point	3,237	3,237 Points
Natural_Cliff	Point	155	155 Points
Natural_Peak	Point	28,106	28,106 Points
Natural_Spring	Point	4,780	4,780 Points
Natural_Tree	Point	533,856	533,856 Points
Natural_Volcano	Point	13	13 Points
Water_Area	Polygon	60,319	2,044,622 Edges
Pois_Area	Polygon	279,010	2,462,611 Edges
Pofw_Area	Polygon	23,337	265,523 Edges
Natural_Area	Polygon	4,034	114184 Edges
Places_Area	Polygon	4,133	273,679 Edges
Landuse_Area	Polygon	459,946	18,528,176 Edges
Buildings_Area	Polygon	2,039,096	14,829,921 Edges

Tab3. Demo Environment

Item	Description
CPU	4core, Intel(R) Xeon(R) CPU E5-2680 v3@2.50GHz
Memory	32 GB
Operating System	Centos7

Application Scenarios

Demo 1 (Buffer analysis of 10-million-scale China datasets)

The 10-million-scale datasets (see Tab 1) used in the demonstration are provided by map service providers. As the datasets are not open published, the raw datasets are encrypted by adding offsets. The interface of the demonstration is simple to use, choose a dataset, input the buffer radius and click the Enter button, then the result layer will be added to the map in real time. Fig 3 shows the analysis results.

Fig 3. Buffer analysis results of China datasets

Demo 2 (Overlay analysis for housing site selection in Spain)

We have designed a housing site selection scenario for VOC based overlay analysis. Suppose that a new immigrant in Spain wants to choose a place to live which meets the following conditions: 1) convenient to traffic (within 500m from Highways); 2) convenient for children education (within 200m from Education amenities); 3) convenient to the medical care (within 2000m from Healthcare amenities); 4) near to leisure places (within 1000m from Entertainment, Arts & Culture amenities or Waterways but not in Water Area); 5) quiet (at least 300m away from Railways). The conditions can be translated into the following expression. Enter the expression and click the Create-Overlay-Layer button (Fig 4), then the result layer will be added to the map in real time. Fig 5 shows the analysis results, in which the red areas are the recommended housing places for the immigrant.

Fig 4. Input of the housing site selection in Spain

Fig 5. Analysis result of the housing site selection in Spain

Contact:

Mengyu Ma@ National University of Defense Technology

Email: mamengyu10@nudt.edu.cn