A Hierarchical Network Simplification via Non-Negative Matrix Factorization

Dias, Markus Diego; Mansour, Moussa Reda; Dias, Fabio; Petronetto, Fabiano; Silva, Claudio Teixeira; Nonato, Luís Gustavo

doi:10.1109/sibgrapi.2017.22

Cited by 10 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since our data is represented as a graph, one option would be the watershed cuts algorithm [36], inspired by the well known image processing segmentation and equally prone to over segmentation. Considering that the processing time is also a relevant factor, we opted for an heuristic variation of the maximum weighted matching algorithm called sorted maximal matching [37], which merges clusters based on the weights of the edges between pairs of clusters.…”

Section: Data Clusteringmentioning

confidence: 99%

“…A graph is generated combining the original census data, encoding the changing geographical information. The graph is partitioned into an hierarchy [37]. The characteristics and evolution of the clusters are then visually represented.…”

Section: Cluster Characterizationmentioning

confidence: 99%

“…Once the distances are associated to the edges, we use watershed cuts [36] to create an initial clustering, which is then refined into a hierarchy using the Sorted Maximal Matching (SMM) [37] with median linkage. The initial watershed step is performed to create an initial clustering and reduce the running time of the SMM.…”

Section: Geographic Content Clusteringmentioning

confidence: 99%

“…The initial watershed step is performed to create an initial clustering and reduce the running time of the SMM. For completeness, we briefly review this method, but we refer the reader to the original paper [37] for more details, including a complete performance evaluation using several metrics.…”

Section: Geographic Content Clusteringmentioning

confidence: 99%

See 3 more Smart Citations

Visualizing demographic evolution using geographically inconsistent census data

Dias¹,

Silver²

2018

Preprint

Self Cite

View full text Add to dashboard Cite

Video Abstract: https://youtu.be/bKeV08Os0uA Census measurements provide reliable demographic data going back centuries. However, their analysis is often hampered by the lack of geographical consistency across time. We propose a visual analytics system that enables the exploration of geographically inconsistent data. Our method also includes incremental developments in the representation, clustering, and visual exploration of census data, allowing an easier understanding of the demographic groups present in a city and their evolution over time. We present the feedback of experts in urban sciences and sociology, along with illustrative scenarios in the USA and Canada.

show abstract

Section: Data Clusteringmentioning

confidence: 99%

Section: Cluster Characterizationmentioning

confidence: 99%

Section: Geographic Content Clusteringmentioning

confidence: 99%

Section: Geographic Content Clusteringmentioning

confidence: 99%

See 2 more Smart Citations

Visualizing demographic evolution using geographically inconsistent census data

Dias¹,

Silver²

2018

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Figure 12 shows two node-link diagrams with different levels of visual clutter. To reduce clutter in node-link diagrams, a number of strategies can be employed, from methods to change nodes' positions, such as force-based and circular algorithms (SIX;TOLLIS, 2006;BATTISTA et al, 1994;MI et al, 2016), to hierarchical simplification (DIAS et al, 2017) and edge-bundling strategies (HOLTEN; WIJK, 2009; LAMBERT; BOURQUI; AUBER, 2010; LHUILLIER; HURTER; TELEA, 2017). Less visual information usually leads to less visual clutter, and so strategies such as node ordering, edge sampling, and temporal resolution change can benefit visual exploration.…”

Section: Visual Clutter In Temporal Network Visualizationmentioning

confidence: 99%

Improving visual analysis of streaming networks

Ponciano¹

View full text Add to dashboard Cite

Temporal networks (also known as dynamic networks) are often used to model connections that occur over time between parts of a system by using nodes and edges. In temporal networks, all nodes, edges, and times, are known and available to be used in the analysis. However, in several real-world applications, data are produced in a massive and continuous way, which is known as data stream. In this case, the volume of data may be so large that the storage may be impossible and mining tasks become more challenging. In streaming temporal networks, edges are continuously arriving in non-stationary distribution. In both temporal and streaming temporal networks, patterns related to node and edge activity are typically irregular in time, which makes the visualization of such networks helpful to gain insights about network structure and dynamics. Nevertheless, the non-stationary distribution of incoming data increases complexity and turns the streaming temporal network visualization even more challenging. Several visualization layouts have been proposed, but they all have limitations. The main challenge in this context is the amount of visual information, that increases depending on the network size and density, and causes visual clutter due to edge overlap, fine temporal resolution, and node proximity. In this thesis, we propose methods to enhance the visualization of streaming temporal networks through the manipulation of the three network dimensions, namely node, edge, and time. Specifically, we propose: (i) CNO, a visual scalable node ordering method; (ii) SEVis, a streaming edge sampling method; and (iii) a streaming method that adapts the temporal resolution according to local levels of node activity. We also present a comparative study considering the combination of these methods. We show through case studies with real-world networks that each of these methods greatly improves layout readability, thus leading to a fast and reliable decision making.

show abstract