Graph Analytics for Avian Science Data

Ami Pandat

1 a

, Minal Bhise

1 b

and Sanjay Srivastava

2 c

Distributed Databases Group, DAIICT, Gandhinagar, India

DAIICT, Gandhinagar, India

Keywords:

Analytics, Centrality, Connectivity, Graph Database, Graph Analytics, Path Analytics.

Abstract:

Data management solutions are becoming increasingly necessary as more Big Data applications are developed.

One such area that deals with Big Data is Big Graphs. Complex relationships exist in graph-based applica-

tions. Analytics and data extraction are better solutions for understanding such complex applications. Data

from Avian Science has shown signiﬁcant growth in recent years. Graph analytics can be used to interpret

complex scientiﬁc data and their relationships. This paper uses graph analytics to discuss the application of

graph analytics in avian science. For the eBird Dataset, four Graph Analytics techniques were identiﬁed and

implemented. These methods extract information about path patterns, node popularity, connections to other

nodes, and clustering. The Dataset includes real-time data on bird observation and distribution. Each analytics

technique extracts data from the birds’ observations. The ﬁndings show that graph analytics for avian science

data can aid in predicting a wide range of crowd-sourced information. Additionally, the work can be expanded

using machine learning methods.

1 INTRODUCTION

Popular modern-day applications, Social Networking,

Transportation Network, Payment-Purchase History,

Crowd-Science, Biodiversity, etc... reports massive

growth in data. To manage this increased data, efﬁ-

cient data management techniques are needed. To ad-

dress the same, there are two approaches: Relational

and Graph-based.

The relational approach for data management han-

dles data in a tabular structure with rows and columns.

Relationships can be formulated through primary and

foreign keys. Complex join operations are required

to retrieve the data through queries from multiple ta-

bles. Several techniques have been identiﬁed for ef-

ﬁcient Query Execution in a relational database for

large datasets: Data Partitioning and allocation (Pan-

dat et al., 2021),(Padiya and Bhise, 2017), data skip-

ping, and Summarization and Distributed Query Pro-

cessing are some of the most popular techniques.

The graph-based approach manages data in the form

of vertex and edge. The Graph is a data structure,

visualized through a triple < vertices, edges , and

relationships between them >. Relationships take

https://orcid.org/0000-0002-6882-9881

https://orcid.org/0000-0003-4364-3930

https://orcid.org/0009-0003-8253-067X

priority in the Graph-based approach. Edges connect-

ing vertices represent the relationship between two

endpoints. The graph-based process eliminates join

operation but suffers scalability and security issues

(Vicknair et al., 2010). Graph partitioning, Graph

summarization (Liu et al., 2016), and Distributed

Graph storage are some solutions for efﬁcient RDF

Data management and query processing.

The systematic computational analysis of data or

statistics, commonly known as analytics, is another

question that data management raises. Analytics de-

tects, evaluates, and conveys essential patterns in

data. Making smarter decisions also involves utilizing

data trends. The structure of the interpretation also

varies amongst analytics. Analytics based on graphs

or tabular/relational data are also possibilities. The

edges connecting the entities are used to carry out the

queries in Graph Analytics(Singh et al., 2018). Com-

pared to a relational database, a graph database exe-

cutes queries more quickly (Vicknair et al., 2010).

The implementation of four types of Graph Ana-

lytics algorithms using Avian scientiﬁc data is shown

in this research. The rest of the article is structured

as follows: The next Section contrasts the standard

Relational and Graph database approaches for avian

research using graph analytics. Section 3 describes

the work connected to graph analytics, and Section

194

Pandat, A., Bhise, M. and Srivastava, S.

Graph Analytics for Avian Science Data.

DOI: 10.5220/0012186000003598

In Proceedings of the 15th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2023) - Volume 3: KMIS, pages 194-201

ISBN: 978-989-758-671-2; ISSN: 2184-3228

4 discusses the Biodiversity domain. Section 5 dis-

cusses the experimental details. Section 6 goes over

how to use the four graph analytics techniques. Sec-

tion 7 is expanded with Hybrid graph analytics tech-

niques, followed by a discussion of the results and

future directions.

2 RELATIONAL AND GRAPH

DATABASES: AVIAN SCIENCE

Comparing both databases concludes that Relational

and Graph-based databases have advantages for dif-

ferent use cases (Vicknair et al., 2010)(Patras et al.,

2021)(Cheng et al., 2019a). A detailed comparison

of the Relational and Graph databases has been pre-

sented in (Cheng et al., 2019b) (Pandat and Bhise,

2022). The primary focus of these analyses is the

Query execution time for four types of queries. The

research states that the Graph database works faster

for Projection and join questions than linear and ag-

gregation queries; relational database performs well.

Let’s compare how relational and Graph databases

handle this speciﬁc use case:

1. Data Model:

• Relational Database: Relational databases are

based on tables with rows and columns. In Avian

Science data, this would mean representing enti-

ties like birds, habitats, and researchers in sepa-

rate tables with relationships established through

foreign keys.

• Graph Database: Graph databases use a graph

data model with nodes (representing entities) and

edges (representing relationships). Each bird,

habitat, researcher, and their connection can be

represented as nodes and edges in a graph.

2. Schema Flexibility:

• Relational Database: Changing the schema to ac-

commodate new relationships or properties can be

challenging and require altering tables, which can

be complex and time-consuming.

• Graph Database: Graph databases are inherently

ﬂexible for handling relationships. You can easily

add new connections or properties to nodes with-

out changing the overall schema.

3. Querying:

• Relational Database: Queries in relational

databases may require complex JOIN operations

to traverse relationships, which can be slower for

deep or complex queries.

• Graph Database: Graph databases excel at travers-

ing relationships, making them well-suited for

graph analytics tasks. Querying for patterns and

relationships is more intuitive and efﬁcient.

4. Performance:

• Relational Database: Relational databases are

generally optimized for structured data with sim-

ple relationships. Complex graph analytics

queries can be slower in relational databases.

• Graph Database: Graph databases are designed

speciﬁcally for graph-related tasks, offering supe-

rior performance for graph analytics.

5. Scalability:

• Relational Database: Scaling a relational database

to handle large-scale graph data can be challeng-

ing and often requires horizontal partitioning or

sharding.

• Graph Database: Graph databases are inherently

scalable for graph-related tasks as they are de-

signed to handle graph data efﬁciently.

6. Use Cases:

• Relational Database: Relational databases are

suitable for data with well-deﬁned, structured re-

lationships and when graph analytics is not the

primary focus.

• Graph Database: Graph databases are ideal when

the relationships within the data are the primary

focus, such as in Avian Science data, where bird

behaviors, interactions, and habitat dependencies

are crucial.

Research states that graph database management sys-

tems still require more security-related features, and

relational databases need more ﬂexibility to adapt to

data changes.

2.1 Database and Ecology

Technically, environmental and ecological data fre-

quently take the form of matrices that may be effec-

tively stored and analyzed using a relational database

management system (RDBMS) or another tabular

data structure.

table

join cost(R, S) = table scan cost(R)

+record(R) × selectivity × records per key(S)

×(CPIO +CPR)

(1)

(Tanaka and Ishikawa, 2019)

When integrating big datasets, the result is fre-

quently kept in volatile memory, a constraint. Table

indices in a standard database design take O(log(n))

Graph Analytics for Avian Science Data

195

time, where n is the size of the input dataset. Reverse

and recursive lookups may be required in a query with

many joins (from various data tables), which might

increase the load from O(n) to O(n

), where k is the

number of data tables to join. Figure 1 explains the

cost of join operation (Tanaka and Ishikawa, 2019)

where CPIO is the I/O cost per page stored record for

DBMS access, and CPR is the CPU cost per record.

3 RELATED WORK

Various algorithms have been devised to discover the

analytical observations from the available data using

the Graph-based approach. These algorithms help

real-life applications efﬁciently analyze the data. This

section summarizes the application of graph analytics

and technical advances observed in Graph Analytics.

Several tools implement the extraction of data using

graph analytics. Many applications, such as perturba-

tion analysis and power failure analysis from graphs

constructed by the power grid, create multiple views

by removing or updating a set of nodes and edges and

then performing computations like path analysis and

so on from scratch inefﬁcient. The motivation behind

the Graphsurge (Sahu and Salihoglu, 2021) system is

a technique that can organize views in a speciﬁc or-

der and carry out the analysis in a manner that can

optimize the time and impair performance.

TurboGraph++ (Ko and Han, 2018) is a scal-

able and fast graph analytics system. It uses

the layered windowed streaming paradigm to con-

duct neighborhood-centric analytics quickly and ef-

ﬁciently with a limited memory budget. The rela-

tional database also implements some of the analytics

based on graph algorithms. Vertexica (Jindal et al.,

2014) is an Analytical tool that performs query exe-

cution in SQL engine for graph queries. It leverages

relational features and uses much more graph anal-

ysis. The popular graph algorithms Page Rank and

Shortest Paths show Vertexica outperforms (Apache,

2011) and regular Graph Database.The recent tuto-

rial presented in (Wang et al., 2020) discusses the

application of Graph Analytics in healthcare, specif-

ically for COVID-19. To deal with large datasets

for Graph Analytics using multi-distributed GPUs has

been presented in (Jatala et al., 2020). It ﬁnds the

evaluation based on four points. (1) The Cartesian

vertex-cut partitioning policy, (2) static load imbal-

ance, (3) device-host communication, and (4) asyn-

chronous execution.

Above all, the framework and operations for Big

data sought a general framework. Recently, in (Bel-

landi et al., 0), a multi-modal Big data analytics

Figure 1: Migration.

framework has been proposed. The proposed work

helps to optimize the cost and helps to improve the

performance of analytical operations. One of the

analytical-based works for the avian science domain

is BioSpytial (Escamilla Molgora et al., 2020). It is

a hybrid analytical tool for crowd science data. It

supports both Relational and Graph-based query pro-

cessing. The work presented in this paper uses the

general framework to implement graph analytics op-

erations. It uses one of the crowd science data eBird

(eBird Database available at:, 2014).

4 AVIAN SCIENCE

Biodiversity data is increasing daily as researchers

and ecologists digitize ecology data. Avian science

is a subset of the vast domain of ecology. Birds and

their observation data help ornithologists and com-

puter scientists analyze their effect on the real world.

Through migration, population, and other factors, an-

alytics of avian science data help researchers in many

ways. Birds are affected majorly by environmental

changes. One such case is Migration. Figure 1 is de-

signed to represent the typical bird migration season.

We can analyze the increasing and decreasing number

of birds in one area over time. The migration of birds

depends not only on time but also on the area they live

in and the availability of resources.

BioSpytial (Biodiversity + Spatial + Python) (Es-

camilla Molgora et al., 2020) aims to discover the co-

occurrence of jaguars with other threatened species

in the borderline of Mexico area. It uses both Raster

and Vector data formats to analyze the Spatial data.

The Graph traversals for 4-neighborhood have been

used to analyze the occurrence of Jaguars in Mexico.

It has been found that 29% Rodents, 23% bats, 15%

deers, and 2% parrots were there in neighboring cells.

There is a clear relation between graph data and eco-

logical/biodiversity data to ﬁnd a network of the same

species in the neighborhood.

This research aims to apply Big Graph Analytics

KMIS 2023 - 15th International Conference on Knowledge Management and Information Systems

196

Figure 2: eBird Graph.

capabilities to determine the statistical impact of the

aforementioned environmental changes on Avian Sci-

ence.

5 EXPERIMENTAL SETUP

This section discusses the experimental setup used

for the evaluation of this work. The Graph Analytics

techniques have been implemented for the following

Dataset utilizing the hardware and software below.

5.1 eBird

eBird (eBird Database available at:, 2014) is a plat-

form Cornell Lab of Ornithology managed to keep

updated data about birds worldwide. eBird is a web-

site collecting bird occurrence and relative abundance

data at speciﬁed English, Spanish, and French places.

Users can choose a place from a drop-down menu

of birding ”hotspots” (shared locations) or utilize

eBird’s online mapping tools to select from or build

new reporting locations when reporting bird sight-

ings. Many participants, for example, designate their

home as a private site and record birds daily, but oth-

ers bird a nearby park every day.

Participants can make repeated observations from

the same area because the chosen locations are saved

in the database. For experiments, two years [2010-

2012] of data for India have been used. The data

size is 80MB, containing 2.3 lacs tuples and 41 at-

tributes. We have prepared a brief structure that ex-

plains the details included in the development of the

Neo4j graph shown in Figure 2. Five entities are cre-

ated for Observers, Birds, Location, State, and Date-

Time details. The Graph shown in Figure 2 is a Prop-

erty graph. Each node contains properties about that

entity. There are four relationships identiﬁed between

available entities. The generated Graph is a directed

graph.

5.2 Hardware and Software

Implementation is done on the system, Intel® Core

(TM) i3-2100 CPU@ 3.10GHz 3.10 GHz 24GB. For

Query execution, Neo4j Desktop 1.1.10 is used, and

Neo4j (Desktop:, 2012) browser version 3.2.19 is

used for visualization. For Analytics purposes, the

Graph data science (GDS) library has been installed

and used for the evaluation. GDS enables hybrid data

analytics. It also needs to be interfaced with Python

language.

6 GRAPH ANALYTICS

Graph analytics include techniques for identifying

strategic entities, uncovering structural data, and cal-

culating data ﬂow in a network. Graph Analytics

helps to analyze and understand the strength between

two nodes using Properties applied to the Graph. For

different use cases, speciﬁc analytics can be used as

an algorithm and gives result relevant to the prod-

uct/company. Regular analytics explores statistics,

computer programming, and operations research to

uncover insights. Graph analytics includes graph-

speciﬁc algorithms to analyze relationships between

entities. Clustering, partitioning, PageRank, and

shortest path algorithms are unique to graph ana-

lytics. One can apply four types of Analytics to

the Graph databases. This section describes all four

types of graph analytics for the ecological Dataset.

We have used the eBird Dataset available at (eBird

Database available at:, 2014).

This motivates a knowledge of its utility and

adaptability for discovery-style analysis for speciﬁc

business problems, their prevalence, and why they are

prevalent. Graph analytics techniques are built on a

model for describing distinct entities and the various

types of relationships that connect them. It uses graph

abstraction to represent connectedness, consisting of

vertices (nodes or points) representing the modeled

items, joined by edges (links, connections, or relation-

ships) that capture how two things are associated.

6.1 Steps to Set up Analytics

Environment

The following steps have been followed to perform

the Big graph analytics on Avian science data.

Graph Analytics for Avian Science Data

197

(a) Original Graph. (b) Clustering Analytics.

Figure 3: Analytics Results.

1. Download the eBird dataset by sending an appli-

cation to eBird developers (it takes around 3-4

working days to get it done)

2. Clean the Dataset based on the attributes you need

for your experiment. (For graph analytics, we

have chosen the attributes mentioned in Figure 2

3. Setup Neo4j Browser and Desktop/Neo4j Aura

based on your system speciﬁcation.

4. For analytics load the labeled in one graph

in Neo4j(for this experiment original generated

graph is presented in Figure 3a)

5. To perform all four types of analysis, create sepa-

rate folders in the Neo4j browser to download the

results and queries

6. Connect the Neo4j browser with Graph Data Sci-

ence(GDS) library

7. Perform your desired analytics experiments as

suggested in the GDS library.

The four Graph Analytics techniques used to perform

avian science data analytics are explained in the fol-

lowing subsections.

6.2 Path Analytics

Path Analytics examines the relationships between

nodes. They are primarily used in shortest-distance

problems. It analyzes similar shapes and distances

from different paths that connect entities within the

Graph.

We can ﬁnd the following details if we map the

eBird dataset to path analytics. 1. The nearest neigh-

bor of the same species. 2. Shortest path between two

common species found in a particular state 3. n-hop

reachability between two vertices.

Cypher Query: MATCH

p=shortestPath((a:BirdNode)-[*]-(c:ObseNode))

Return p, length(p) LIMIT 25

6.3 Connectivity Analytics

This type of Analytics determines how strongly or

weakly connected two nodes are. Connectivity anal-

ysis outlines the number of edges ﬂowing into the

node and those ﬂowing out. This analysis provides

a method to identify malicious or unexpected patterns

within the data. It gives the best solution to ﬁnding

connectivity between different entities.

Graph Analytics based on connectivity helps to

ﬁnd a connection between observers and bird species.

We can apply the same to ﬁnd the below-mentioned

query solution.

1. Number of observers who observed particular

species

2. Number of species observed by a particular ob-

server

3. Number of observers at a particular landmark

Cypher Query: Match (n:BirdNode)-[r]-() return

n.Name, count(distinct r) as degree Order by degree

6.4 Centrality Analytics

Estimates how important a node is for the connectiv-

ity of the network. Using the PageRank algorithm

helps to estimate the most inﬂuential people in a so-

cial network or most frequently accessed web pages.

It helps to evaluate the importance of a present

node within the graph network and its connectivity to

others. If one would like to ﬁnd the most inﬂuential

node, this is the technique. For eBird, we can ﬁnd the

KMIS 2023 - 15th International Conference on Knowledge Management and Information Systems

198

Figure 4: Community Analytics.

most visited place or species by an observer in all the

seasons. It will help to put analysis in a time-ordered

manner.

6.5 Community Analytics

Distance and density of relationships can be used to

ﬁnd groups of people frequently interacting in a social

network. Community analytics also deals with the de-

tection and behavior patterns of communities. Using

graph analytics saves time. Graph analytics are eas-

ier to work with than the traditional techniques being

used. Modeling the data and its storage becomes easy.

Patterns can aid data-driven decision-making when

appropriately understood with the correct meaning

derived from the data. Overloaded and strained re-

sources within the organization can be identiﬁed and

reconﬁgured using graph analytics. The more it is

connected, the most important it is in the network.

The community analysis helps to ﬁnd the most fre-

quently observed bird species from the network.

Cypher Query: MATCH (n:ObseNode)-[r]-

¿(m:BirdNode) WHERE n.Name

=”obsr360080” RETURN n,r,m

7 HYBRID ANALYTICS

In cases where interpretation must be made from one

or more analytics, hybrid analytics can be performed.

For the same, we have conducted two hybrid analyt-

ics.

7.1 Path-Cluster Analytics

In avian science, we are often required to ﬁnd a ﬂock

of birds moving or migrating in certain paths. Clus-

Figure 5: Path Analytics.

tering analytics results in clusters of species. To ﬁnd

those clusters’ source and target destination, the clus-

ters would be fed to the path analytics query, and the

moving path can be analyzed.

Cypher Query: Find a path of clusters of Yellow-

browed Bulbul between any two locations.

7.2 Community-Cluster Analytics

The degree of nodes, i.e., Community analytics, an-

swers the number of nodes connected to several

nodes. However, the issue arises when we would like

to ﬁnd only the degree of nodes in clusters. For that,

we need ﬁrst to form a cluster, and then one can count

the degree in a cluster of nodes. This hybrid analyt-

ics helps to answer some of the following queries of

Avian Science data.

Cypher Query: Find a number of locations where

Slaty-blue Flycatcher is found in a ﬂock of 100.

8 ANALYTICS RESULTS AND

DISCUSSION

This section discusses the results from the experi-

ments performed for all four types of analytics. Fig-

ure 3b shows the clusters formed from the original

Graph. The clustering analytics help to retrieve the

information that can be clustered together. The ﬁg-

ure shows the two clusters generated from the original

Graph in Figure 3a. The clusters help to identify the

connectedness of similar entities. Community analyt-

ics is just the extended version of clustering analytics.

Clusters formed based on the particular community

can be identiﬁed, and as shown in Figure 4, a dif-

ferent community of bird species and observers are

represented. Numerous species are available in the

Dataset; each can be identiﬁed as a community. This

Graph Analytics for Avian Science Data

199

(a) Link Prediction. (b) Node Prediction.

Figure 6: Machine Learning Approaches.

Table 1: Centrality Analytics.

Species Centrality

Greenish Warbler 0.78

Slaty-blue Flycatcher 0.23

Yellow-browed Bulbul 0.67

Oriental Honey-buzzard 0.37

Black Kite 1

type of analytics helps to determine the similar data

available in the Dataset.

Path Analytics helps to analyze the path-related

properties between two entities. We can determine

particular entities’ shortest, largest, and nearest neigh-

bors. Figure 5 shows the available path between bird

species found in nearby areas. This analytics can be

applied to ﬁnd the n-neighborhood of the entities.

The centrality analytics help to identify the most

popular nodes in the network. The value of centrality

varies between 0 to 1. The value 1 shows the most

central node in the network. The table shows the cen-

trality values for the bird node. Four types can repre-

sent centrality analytics. Degree centrality analytics

is one of them. The result shown in the table is for

degree centrality analytics.

8.1 Challenges

Numerous characteristics of graph problems provide

severe obstacles to effective parallelism. High-degree

vertices in graphs are prevalent. These graphs could

be more computationally burdensome to partition. In

real-world graphs, most vertices have comparatively

few neighbors, while a few have numerous neighbors.

More computation, coordination, and communication

are needed to partition the sparse graphs.

Natural graphs have colossal sizes that are too

large to ﬁt in a single memory. There are commu-

nication costs as a result of the high-degree vertices.

9 MACHINE LEARNING

TECHNIQUES FOR GRAPH

ANALYTICS

Graph analytics applications aren’t just for interpret-

ing data; they can also predict how data will change

shortly. Graph analytics applications can use data

from recent years to forecast future changes in the

centrality and community of bird data.

As shown in Figure 6a and 6b, we can predict the

missing observation data from the eBird dataset. In

observation, there may be a case where nodes or links

are missing in the data. Machine Learning (ML) al-

gorithms help to ﬁnd the same. We can identify and

categorize the birds based on their characteristics us-

ing classiﬁcation and graph analytics techniques. For

example, some birds are autumn birds in one location

but spring birds in another. Machine learning solu-

tions can be used to perform these classiﬁcations. The

rate of change in bird observation data and noise in

the data can be analyzed using a regression approach.

Various open issues in avian science can be addressed

by combining machine learning and graph analytics.

10 CONCLUSION

Big Graph Analytics techniques for Avian Science

data have been implemented in this paper. Popu-

lar bird observation data eBird has been used to per-

form the experiments. The analytics experiments help

identify the dependence between observers and birds,

including locality parameters. Community analytics

form clusters of identical communities; connectivity

shows the strongly connected component with num-

bers in the form of in-degree and out-degree, whereas

centrality analytics help to identify the most popu-

lar entity, and path analytics determine the path be-

tween the source and target entity. When dealing with

KMIS 2023 - 15th International Conference on Knowledge Management and Information Systems

200

Avian Science data and performing analytics, a graph

database is generally a better choice due to its ﬂex-

ibility, performance, and inherent support for han-

dling complex relationships within the data. Rela-

tional databases can work for this use case but may

be less efﬁcient and more complicated to model and

query. Hybrid analytics techniques help determine

the signiﬁcance of multiple analytics on a domain.

Further, ML techniques help to predict the relation-

ship/link between different entities.

REFERENCES

Apache (2011). Apache Giraph Accessible at:https://giraph.

apache.org/:.

Bellandi, V., Ceravolo, P., Maghool, S., and Siccardi, S. (0).

Toward a general framework for multimodal big data

analysis. Big Data, 0(0):null. PMID: 35666602.

Cheng, Y., Ding, P., Wang, T., Lu, W., and Du, X. (2019a).

Which category is better: Benchmarking relational

and graph database management systems. Data Sci-

ence and Engineering, 4(4):309–322.

Cheng, Y., Ding, P., Wang, T., Lu, W., and Du, X. (2019b).

Which category is better: Benchmarking relational

and graph database management systems. Data Sci-

ence and Engineering, 4(4):309–322.

Desktop:, N. (2012). {https://neo4j.com/},.

eBird Database available at: (2014). https://ebird.org/home.

Escamilla Molgora, J. M., Sedda, L., and Atkinson, P. M.

(2020). Biospytial: spatial graph-based computing for

ecological Big Data. GigaScience, 9(5). giaa039.

Jatala, V., Dathathri, R., Gill, G., Hoang, L., Nandivada,

V. K., and Pingali, K. (2020). A study of graph ana-

lytics for massive datasets on distributed multi-gpus.

In 2020 IEEE International Parallel and Distributed

Processing Symposium (IPDPS), New Orleans, LA,

USA, May 18-22, 2020, pages 84–94. IEEE.

Jindal, A., Rawlani, P., Wu, E., Madden, S., Deshpande, A.,

and Stonebraker, M. (2014). VERTEXICA: your rela-

tional friend for graph analytics! Proc. VLDB Endow.,

7(13):1669–1672.

Ko, S. and Han, W.-S. (2018). Turbograph++: A scalable

and fast graph analytics system. In Proceedings of

the 2018 International Conference on Management of

Data, SIGMOD ’18, page 395–410, New York, NY,

USA. Association for Computing Machinery.

Liu, Y., Dighe, A., Safavi, T., and Koutra, D. (2016).

A graph summarization: A survey. CoRR,

abs/1612.04883.

Padiya, T. and Bhise, M. (2017). DWAHP: workload aware

hybrid partitioning and distribution of RDF data. In

Desai, B. C., Hong, J., and McClatchey, R., editors,

Proceedings of the 21st International Database En-

gineering & Applications Symposium, IDEAS 2017,

Bristol, United Kingdom, July 12-14, 2017, pages

235–241. ACM.

Pandat, A. and Bhise, M. (2022). Rdf query processing:

Relational vs. graph approach. In Singh, P. K., Wierz-

cho

n, S. T., Chhabra, J. K., and Tanwar, S., editors,

Futuristic Trends in Networks and Computing Tech-

nologies, pages 575–587, Singapore. Springer Nature

Singapore.

Pandat, A., Gupta, N., and Bhise, M. (2021). Load balanced

semantic aware distributed RDF graph. In IDEAS

2021: 25th International Database Engineering &

Applications Symposium, Montreal, QC, Canada, July

14-16, 2021, pages 127–133. ACM.

Patras, V., Laskas, P., Koritsoglou, K., Fudos, I., and Kar-

vounis, E. (2021). A comparative evaluation of rdbms

and gdbms for shortest path operations on pedestrian

navigation data. In 2021 6th South-East Europe De-

sign Automation, Computer Engineering, Computer

Networks and Social Media Conference (SEEDA-

CECNSM), pages 1–5.

Sahu, S. and Salihoglu, S. (2021). Graphsurge: Graph

analytics on view collections using differential com-

putation. In Proceedings of the 2021 International

Conference on Management of Data, SIGMOD ’21,

page 1518–1530, New York, NY, USA. Association

for Computing Machinery.

Singh, D., Dutta Pramanik, P., and Choudhury, P. (2018).

Big Graph Analytics: Techniques, Tools, Challenges,

and Applications, pages 195–222.

Tanaka, T. and Ishikawa, H. (2019). Measurement-based

cost calculation method focusing on cpu architecture

for database query optimization. In Proceedings of

the 11th International Conference on Management of

Digital EcoSystems, MEDES ’19, page 56–65, New

York, NY, USA. Association for Computing Machin-

ery.

Vicknair, C., Macias, M., Zhao, Z., Nan, X., Chen, Y., and

Wilkins, D. (2010). A comparison of a graph database

and a relational database: A data provenance perspec-

tive. In Proceedings of the 48th Annual Southeast

Regional Conference, ACM SE ’10, New York, NY,

USA. Association for Computing Machinery.

Wang, F., Cui, P., Pei, J., Song, Y., and Zang, C. (2020).

Recent advances on graph analytics and its appli-

cations in healthcare. In Proceedings of the 26th

ACM SIGKDD International Conference on Knowl-

edge Discovery and Data Mining, KDD ’20, page

3545–3546, New York, NY, USA. Association for

Computing Machinery.

Graph Analytics for Avian Science Data

201