Exploration and Analysis of Temporal Property Graphs Christopher Rost University of Leipzig & ScaDS.AI Dresden/Leipzig, Germany rost@informatik.uni-leipzig.de Kevin Gomez University of Leipzig & ScaDS.AI Dresden/Leipzig, Germany gomez@informatik.uni-leipzig.de Philip Fritzsche University of Leipzig & ScaDS.AI Dresden/Leipzig, Germany fritzsche@informatik.uni-leipzig.de Andreas Thor Leipzig University of Applied Sciences, Germany andreas.thor@htwk-leipzig.de Erhard Rahm University of Leipzig & ScaDS.AI Dresden/Leipzig, Germany rahm@informatik.uni-leipzig.de ABSTRACT We demonstrate the Temporal Graph Explorer, a distributed open- source framework that enables time-dependent graph exploration and analysis on large real-world networks using a rich tempo- ral property graph model and dynamic graph operators. In this demonstration we show how the evolution of a graph plays an essential role in many analytical questions. Besides retrieving a snapshot from a past graph state or calculating the diference between two graph snapshots, users can use our application to summarize the graph to reduce its complexity and to obtain deeper insights into its evolution. Visitors of the demo can vi- sually experience these advanced temporal operators and their results or build and manipulate analytical programs in a declar- ative way without extended programming skills and run them on a distributed local or remote environment. We provide real- world temporal graph data from bicycle rentals of New York City with millions of rentals per month, also demonstrating horizontal scalability of the system. 1 INTRODUCTION Graphs are an intuitive way to model and analyze complex re- lationships between entities representing real-world scenarios. Since most entities and interconnections evolve in the real-world, graphs also change over time in terms of their structure and con- tent. For example, Figure 1 shows a toy example of bicycle rentals (represented as directed edges) between fxed stations (vertices) over time. Such temporal property graphs [1] additionally allow tracking changes in the graph over time. In the example of Fig- ure 1 both vertices and edges store temporal information (marked by a clock symbol) as attributes (valid times) and, thus, the graph reveals that the bike with id 2115 was moved from station [1] to [2] by three consecutive rentals over time. In this paper, we demonstrate the Temporal Graph Explorer, a tool for user-friendly exploration, analysis, and visualization of large temporal property graphs. The core of this application is Gradoop 1 , an open-source framework for distributed graph analysis. Its Temporal Property Graph Model (TPGM) [8] enables modeling and analysis of graphs with bitemporal time seman- tics. The TPGM also comes with a set of composable temporal graph operators (including snapshot retrieval, graph evolution, and time-dependent grouping and aggregation) that can be visu- ally confgured through the user interface or programmatically 1 https://github.com/dbs-leipzig/gradoop © 2021 Copyright held by the owner/author(s). Published in Proceedings of the 24th International Conference on Extending Database Technology (EDBT), March 23-26, 2021, ISBN 978-3-89318-084-4 on OpenProceedings.org. Distribution of this paper is permitted under the terms of the Creative Commons license CC-by-nc-nd 4.0. Figure 1: Example temporal property graph representing bicycle rentals between rental stations. The validity pe- riod of an edge is marked with a clock symbol and sim- plifed with numbers instead of timestamps. combined with the help of the declarative language GrALa [5] to build distributed analysis workfows. The Temporal Graph Explorer thus enables the analysis of the evolution of graphs, i.e., to fgure out when something happened or changed, rather than a static view representing something that happened at some time [9]. To this end, it provides temporal graph operators (including snapshot retrieval and graph difer- ence) to compute and visualize changes, including additions and deletions, that have been occurred. This can be used in the given bicycle example to fnd, for a given week in the past, all rentals that have been added, removed, or remained the same. Our Temporal Graph Explorer can also be employed for the analysis of large graphs by using time-related grouping and aggregation. This allows for a profound exploration of a graph’s structure, semantics, and development over time, which is a sig- nifcant part of knowledge discovery for temporal graphs. Such a graph grouping mechanism helps to fnd out how diferent types of vertices and edges are connected as well as when and how long they were connected. In addition, the graph can be grouped on diferent dimensions, e.g., by rental time or by station location, as well as on diferent dimension levels, e.g., per year or month for the time dimension. The grouped vertices and edges can fur- ther be aggregated in any conceivable way, from a simple count to the minimum, maximum and average duration of a specifc relationship type. 2 TEMPORAL GRAPH ANALYSIS AND EXPLORATION The Temporal Graph Explorer is an application to explore, analyze and visualize temporal property graphs. An intuitive web-based user interface enables the confguration of selected temporal Demo Series ISSN: 2367-2005 682 10.5441/002/edbt.2021.83