jeremychou28 / geoviz

GeoViz: A Multi-View Visual Platform for Spatio-temporal Knowledge Graph

GeoViz: A Multi-View Visual Platform for Spatio-temporal Knowledge Graph

demo video

Overview

We propose a Tree-Net-Map multi-view visualization method to assist in analyzing STKG. Specifically, we have customized three visualization modes for STKG, namely knowledge tree, knowledge net, and knowledge map. Our approach not only reflects the rich attributes and relationships between knowledge triplets, but also highlights the inherent spatio-temporal characteristics of knowledge triplets from three distinct perspectives. For the convenience of users, we have developed a one-stop platform called GeoViz to support customized visualization needs with an easy-to-use web user interface. The high level system design of our GeoViz is shown in Figure. Demo visitors can explore our services via https://mdkg.acemap.cn/#/service.

For a better user experience, GeoViz will continuous to update.

How to use GeoViz?

Workflow

Figure 1 shows the workflow of our platform.

1. Access the homepage (A) of our platform through the web user interface, which consists of two parts: visualization and knowledge discovery.
2. Upload users’ spatio-temporal knowledge graph data based on customized service requirements, and then define relevant settings (B) for the back-end to parse the received data.
3. After successful completion of the preceding steps, users can receive a link to access visualization results of three distinct modes: the knowledge tree (C1), the knowledge net (C2), and the knowledge map (C3).
4. Choose the knowledge discovery part (D1), which utilizes the pre-defined prompt in the back-end to discover novel knowledge links from existing data (D2).

Demo

Demo data

Users can upload their own spatio-temporal knowledge graph (STKG) data to complete their customization needs.

Here we give a demo data for case study.

demo_data/node_data.csv is the entity data in STKG.

demo_data/edge_data.csv is the relationship data between entities in STKG.

Discussion

Tools	Tree-Net-Map(multi-view)	ST Filter	ST Interaction	Knowledge Discovery	Open Source
D3.js	×	×	×	×	√
Vis.js	×	×	×	×	√
Echarts	×	×	×	×	√
Antv G6	×	×	×	×	√
Neo4j(Database)	×	√	√	×	√
GeoViz	√	√	√	√	√

GeoViz is the first concentrated on Tree-Net-Map multi-view visual platform for STKG. Moreover, GeoViz has spatio-temporal filters(ST Filter) and supports dynamic spatio-temporal interactions(ST Interaction). In addition, GeoViz also provides the function of autonomous knowledge discovery. Although Neo4j provides a web user interface for knowledge graph visualization, it requires interaction through the Cypher programming language, which is unfriendly to users. Last but not least, GeoViz, like mainstream visualization tools, is open-source.