Big Data Technology Architecture Proposal for Smart Agriculture for

Moroccan Fish Farming

SARAH BENJELLOUN, MOHAMED EL MEHDI EL AISSI,

YOUNES LAKHRISSI, SAFAE EL HAJ BEN ALI

University of Sidi Mohamed Ben Abdellah Fes,

MOROCCO

Abstract: - As the global population increases rapidly, so does the need for fishing products. Aquaculture is

well-developed in Asian countries but is underdeveloped in countries that share Morocco's climate. To meet the

rising demands for aquaculture production, it is vital to embrace new digital strategies to manage the massive

amount of data generated by the aquaculture environment. By employing Big Data methodologies, aquaculture

activity is handled more effectively, resulting in increased production and decreased waste. This phase enables

fish farmers and academics to obtain valuable data, increasing their productivity. Although Big Data

approaches provide numerous benefits, they have yet to be substantially implemented in agriculture,

particularly in fish farming. Numerous research projects investigate the use of Big Data in agriculture, but only

some offer light on the applicability of these technologies to fish farming. In addition, no research has yet been

undertaken for the Moroccan use case. This study aims to demonstrate the significance of investing in

aquaculture powered by Big Data. This study provides data on the situation of aquaculture in Morocco in order

to identify areas for improvement. The paper then describes the adoption of Big Data technology to intelligent

fish farming and proposes a dedicated architecture to address the feasibility of the solution. In addition,

methodologies for data collecting, data processing, and analytics are highlighted. This article illuminates the

possibilities of Big Data in the aquaculture business. It demonstrates the technological and functional necessity

of incorporating Big Data into traditional fish farming methods. Following this, a concept for an intelligent fish

farming system based on Big Data technology is presented.

Keywords: - Data-driven Fish Farming, Big Data Management, Data Lake, Data Processing, Data Insights,

Smart Aquaculture.

Received: March 27, 2022. Revised: October 22, 2022. Accepted: November 25, 2022. Published: December 16, 2022.

1 Introduction

All spheres of endeavor have experienced

substantial progress thanks to the growth of

robotics, the Internet of Things (IoT), fifth-

generation (5G), Big Data, and artificial

intelligence, [1]. These technologies accelerate the

emergence of intelligent industries. Agriculture is

the world's most crucial industry, [2]. The demand

for food increases along with the global

population, [3]. Fish is a form of protein that is

leaner and lower in calories. Consequently, fish are

extensively consumed, [4].

The contribution of aquaculture to global food

safety is essential. In recent years, aquaculture

production has been expanding all around the

world. According to The World Bank, fish

production climbed in fifty years from 1 million

tons in the 50s to 55 million tons in 2004 to 90

million tons in 2012 and reached 106 million tons

in 2015, [5]. Regarding fish output, Asia ranks

first, with China as the most significant

aquaculture producer, followed by Indonesia,

India, Vietnam, the Philippines, Bangladesh, South

Korea, Thailand, and Japan. Unfortunately,

Morocco needs to improve its performance in this

area. Aquaculture in Africa, in general, is limited

despite huge prospects, [6]. Morocco is our

research location since it is the country of

residence and affords contact convenience.

Big Data offers tremendous promise for

evaluating agricultural and aquacultural

productivity-enhancing interventions. Indeed, all

industries have benefited greatly from the

advantages of data-oriented solution and their

implementations. According to [7], in 2019, the

total of data analysis and business development

reached 189.1 billion US dollars; in 2022, this

amount is estimated to be 274.3 billion US dollars.

These statistics urge the adoption of data solutions

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

311

Volume 19, 2022

to achieve their benefits. Adopting Data-oriented

solutions in fish farming permits intelligent

management and smart decision-making. Big Data

solutions aid sectors and businesses in decision-

making by handling and analyzing massive

amounts of data. Diverse organizations invest in

Big Data technologies to uncover hidden patterns,

market trends, client preferences, and unidentified

linkages.

Big Data solutions have been applied in

numerous areas, such as banking, insurance,

medicine, industry, and marketing. Even though

Big Data has gained much success in the stated

disciplines, it began being adopted in agriculture

only lately, even though not in the fish farming

domain, [8]. Agriculture was the first industry to

benefit from Big Data-related technology [10],

followed by a few aquaculture commercial

solutions. Based on agriculture specialists'

feedback, adopting Big Data to fish farming could

enhance profits and the quality of products, [9].

For this significant purpose, implementing new

Big Data approaches has become vital to face the

difficulties of productivity, environmental impact,

food security, and sustainability. The inspiration

for producing this paper originates from the fact

that Big Data is a relatively new approach

underutilized in fish farming despite its proven

benefits in other fields.

The primary objective of this contribution is to

demonstrate how Big Data can overcome obstacles

in fish farming by providing a Big Data

Architecture for fish farming systems.

In this regard, the current research intends to

offer a functional architecture and expand it to a

Big Data-based technical architecture. This effort

aims to develop a system that not only manages

but also optimizes the production of fish by

utilizing already-existing data. The idea is initially

to get data, process it, and store it before

employing it for reports, dashboards, and

predictive purposes.

This report provides data on the current

situation of aquaculture in Morocco to demonstrate

the need for improvements in this field. Then, we

illustrate the various applications of Big Data in

this sector by analyzing research publications on

data-driven investigations. Then, we concentrate

on the various Big Data strategies that have the

potential to improve fish farming production.

Finally, both functional and technological

architecture proposals are presented.

The structure of this work consists of seven

primary components. The first section outlines the

condition of aquaculture in Morocco. The

following part describes the procedure we took to

conduct the research. In the third section, we stress

use cases of Big Data connected to fish farming.

The three key use cases are management,

optimization, and prediction. The fourth segment

offers Big Data approaches for Fish Farming. In

the fifth section, we propose a functional

architecture for Big Data for the fish farming use

case. The sixth section outlines each component of

a dedicated technical architecture proposal. The

seventh section comprises works that are linked.

2 Aquaculture in Morocco

Aquaculture in Morocco continues to expand at a

reasonably modest rate compared to other locations

worldwide. Currently, Asian countries produce the

highest amounts of aquaculture products.

According to The World Bank [11], Morocco's

continental aquaculture production climbed from

1,403 tons in 2001 to 2,250 tons in 2005. The

production level decreased to 579 tons in 2008

before rising to 1,267 tons in 2018. (figure 1).

According to [12], annual aquaculture

production has reached 13,000 tons, mainly from

reservoirs, lakes, and rivers. The residual

production of continental aquaculture in 2018 is

distributed as follows:

 Eel (estimated production level of 350

tons/year),

 Tilapia (about 200 tons/year),

 Trout (100 tons/year),

 An unspecified amount of production from

reservoir fisheries of carp and other species.

Therefore, it produced a total of 14,267 tons of

fish in 2018, and according to Our World in Data,

the per capita consumption of fish and seafood in

Morocco was 19.47 kilograms, for a total of

692,742,6 tons.

Given the limited production, it cannot be

denied that the aquaculture industry in Morocco is

still in its infancy. In addition, according to the

International Trade Center's (ITC) Trade Map [13],

the value of Morocco's fish imports is 216 032

million US dollars.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

312

Volume 19, 2022

Fig. 1: Moroccan Aquaculture production (metric tons) 1991-2018, [11].

Fig. 2: Moroccan consumption of Fisheries per capita 1991-2017, [14].

In addition, the fish and seafood consumption

per capita in Morocco in 2017 was 19.47kg, which

equates to a total of 692,742.6 tons of fish and

seafood consumed in Morocco (figure 2).

When comparing the fish production (14.267

tons) and fish consumption (692,742,6 tons), a

significant deficit must be closed over time,

necessitating the importation of fish products. In

order to meet the rising demand, it is necessary to

replace traditional fish farming systems with

intelligent fish farming systems. The primary

purpose of this work is to reveal the potential of

Big Data Technology and how it might transform

the quality and management of fish production

through a data-driven fish farming system.

3 Methodology

To conduct the study, we explored existing

methodologies using a conventional approach. All

available literature released after 2015 has been

evaluated. In addition to the period criterion, we

employed two inclusion criteria: publishing the

complete paper and relevance to the research. In

addition, two exclusion criteria were utilized:

English-language language publications and

articles focusing on technical design. We used the

following query to retrieve articles from updated

research databases such as Web of Science,

Springer, and IEEE Xplore: ["Big Data" OR "Data-

driven" OR "Big Data Technologies" OR "Data

Lake Architecture"] AND ["Aquaculture" OR

500

1000

1500

2000

2500

3000

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

Aquaculture production (metric tons) - Morocco

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

Fish and seafood consumption per capita (Kg) -

Morocco

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

313

Volume 19, 2022

"Fish Farming"] AND "Morocco."

Unfortunately, no publications about Smart

fish farming utilizing Big Data in Morocco could

be identified. Instead, we found the majority of

publications on Asia and specifically used China.

In order to solve this difficulty of insufficient

references, we based our research on articles from

countries with climates nearly identical to Spain. In

addition to Spain, no pertinent publications from

Algeria or Tunisia were discovered.

After gathering the articles, our methodology

consisted of the subsequent steps: All relevant Big

Data-based aquaculture papers were gathered;

since Asia is the leader in aquaculture production,

articles from China were selected to examine the

various methodologies employed in this region.

We limited our research by focusing on applying

certain techniques in Spain, which shares a climate

with Morocco.

4 Big Data in Smart Fish Farming

Use Cases

Big Data is one of the pillars of the 4.0 industry,

and it has been adopted in numerous industries,

including those in the following: healthcare,

manufacturing, entertainment, cyber security and

intelligence, crime prediction and prevention,

science, traffic optimization, and weather

forecasting. Big Data gives firms invaluable

insights and undeniable profits, [15], [16].

As smart technologies and sensors proliferate

on farms and the quantity and scope of farm data

expand, farming processes will become

increasingly data-driven and data-enabled. Rapid

Internet of Things and Cloud Computing advances

drive the Smart Farming phenomena, [17].

These technologies can be utilized in fish

farming to forecast patterns, boost productivity and

profitability, and enhance fish quality. Some

researchers are interested in Big Data for its

potential to contribute to the sustainability of

aquaculture, [18], [8].

Fish farms continuously generate and gather

data [19]. The aquaculture data value chain begins

with data acquisition from either streaming or

batch data sources [20]. Preprocessing is used to

sanitize collected data for validation reasons. This

data is subsequently stored in a distributed storage

system so that Data Analysis can be performed,

[15]. There are four distinct Data Analysis

categories:

● Descriptive Analysis: Utilize dashboards to

track Key Performance Indicators (KPIs);

● Diagnostic Analysis: Apply drill-down of

descriptive analysis to identify patterns of

behavior;

● Predictive Analysis: Use statistical modeling to

forecast what is most likely to occur;

● Prescriptive Analysis: Use the learnings from

all prior analyses to decide what actions to

take.

Data Visualization is the graphical depiction of

data, which may be utilized to monitor IoT devices'

production and activity status using data supplied

by these devices.

Incorporating Big Data technology into fish

farms aims to improve production per cost quota

and environmental footprint by enhancing fish

survival and water quality for greater

sustainability, [19], [21]. Growth, survival, and

feed conversion ratio (FCR) are aquaculture's most

economically essential qualities, [22]. Numerous

articles examine these issues to comprehend the

circumstances that influence these three

characteristics.

As China is the global leader in the

aquaculture industry, numerous research articles

utilize Big Data to examine aquaculture. Relevant

applications discovered can be categorized into six

groups [23]: live fish identification, species

classification, behavioral analysis, feeding

decisions, size or biomass calculation, and water

quality prediction. In addition to focusing on

economic factors, preserving an ecological

environment with high water quality is believed to

be the most crucial factor in ensuring production

efficiency with high quality, [24]. Consequently,

significant research has been undertaken in water

quality management utilizing Big Data

applications, [24], [25], [26], [27], [28].

Comparatively, Spain's research publications focus

on fish feeding strategies and tanks' water quality

management for enhancing the operational process'

economic efficiency, [29], [30], [31].

5 Big Data Techniques for Smart Fish

Farming

Technology is one of the effective ways to boost

agricultural productivity in terms of quality and

quantity. Big data technology and practices

effectively enhance agricultural production and

tackle future difficulties. In general, two concepts

of database management systems can be

distinguished.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

314

Volume 19, 2022

5.1 Data Warehouse

The Datawarehouse architecture uses a relational

database to allow data analysis and reporting.

Typically, it consists of structured historical data

from transactional databases, [32]. In addition,

using the data warehouse concept, a subject-

oriented approach is adopted to store the data,

meaning that each table is constructed to precisely

meet a previously determined need. Moreover,

before saving data in the landing zone of a DWH,

it must adhere to a specific structure. In other

terms, it may undergo data cleansing operations to

ensure excellent data quality for reporting, [33].

In smart fish farming, data and information are

the fundamental components. The collection and

sophisticated analytics of all or a portion of the

data will enable decision-making grounded in

science. Nonetheless, the vast quantity of data in

smart fish farming presents numerous obstacles,

including different sources, varied formats, and

complex data, [34].

Information about the devices, the fish and its

milieu, and the breeding process are available from

numerous sources in various formats such as

pictures, audio, and text files. The complexity of

the data stems from the various species, modes,

and cultivation phases. Managing those mentioned

above massive, nonlinear, and high-dimensional

data is a challenging undertaking. In addition, the

ETL (Extract, Transform, Load) is the approach to

collecting data in a data warehouse architecture.

The extraction step represents extracting

heterogeneous/homogeneous data from the source;

the second step is transforming data into an

appropriate form and cleaning it to make querying

and analysis more efficient. Data loading refers to

storing data in the target.

5.2 Data Lake

The data lake architecture is characterized as a

robust data infrastructure for storing enormous

quantities of data, characterized by its variety. In

addition, it offers to store each data type in its raw

format, despite its origin, [35].

The Data Lake provides a high degree of

flexibility since the captured data is schema-less;

all the data can be stored regardless of the design

or the requirement to know the future use case and

where our acquired data should deliver answers,

[36]. This architecture provides excellent

flexibility when analyzing data through a querying

language, Data analytics, full-text searching, real-

time analytics, and machine learning.

In a data lake, we refer to an ELT (Extract,

Transform, and Load) process rather than an ETL

(Extract, Transform, and Load), which is the

fundamental data-acquiring method for the data

warehouses. In this perspective, a data lake

architecture's priority is collecting and storing data

in its file system to constitute a solid historical

database, [33]. After that, the transformation phase

is where we apply the preprocessing of data to

build data suitable for each use case.

In the following table (table 1), we present a

comparative overview between the Data Lake

architecture and the Data warehouse architecture:

Table 1. Data Warehouse versus Data Lake.

Characteristics

Data Warehouse

Data Lake

Data

Relational from Transactional

systems and operational

databases

Relational/non-relational from IoT

devices, social media

Schema

On-write

On-read

Users

Business Analysts

Data Scientists, Data Engineers and

Business Analysts.

Analytics

Batch reporting, B.I., and

visualization

Machine Learning, Predictive

analytics, data discovery, and

profiling

To choose the appropriate architecture, it is

essential to note that the collected fish farming data

differs from the data previously kept in standard

database management systems because it consists

of huge amounts of data.

In addition, we must determine if the data

created by fish farms can be categorized as Big

Data in order to select the most appropriate

architecture for data management. Due to its

qualities, this information may not qualify as Big

Data. Nevertheless, some researchers believe it to

be Big Data and note that it will be Big Data if all

data from the fish farming system are compiled,

[8], [36]. The following section describes the

properties of fish farm data that come inside the 5

Vs of Big Data, [37].

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

315

Volume 19, 2022

● Volume: Fish data is often produced by

various devices, including tanks, pumps,

manual measurements, and sensors, [38].

Usually, workplace PCs or even cloud services

are used to store this data. According to

statistics, fish farming systems produce and

retain enormous volumes of data organized by

particular years. Because of this, it can be

challenging to transfer large amounts of data to

other devices.

● Velocity: Velocity represents the rapid changes

in data characteristics. The average hourly data

output from various sensors is 10 MB. The

data size keeps growing depending on the

actions involved in fish production, [39]. As a

result, data size grows when fish output is at its

peak but continues even when production is

low. Fish have different lifespans that vary

from one to the next. As a result, fish data

differ from one another.

● Variety: Variety is the different types of data

from multiple sources. The variety is wide

when data is structured, semi-structured, and

unstructured. Data from fish farms should be

organized based on where it came from to

comprehend better and utilize the information.

Data can occasionally differ from device to

device, from an automatic sensor to a human

technique, [39], [40]. Data should be managed

appropriately according to its collection, types,

and procedures. It is crucial to appropriately

input manual data into the computer and

integrate it with machine-based sensors, [41].

It is necessary to analyze and organize the

manually gathered data into a structured

format for later usage.

● Veracity: In fish farming, manually generated

data generally is unstructured, [41]. The

quality of the data is also an issue. Manually

obtained data are typically noisier than data

from machine-based sensors, [42].

Additionally, sensor elements and human

factors significantly impact fish farming data.

These influences can be reduced by keeping

accurate records of manually applied inputs

and adding automated sensors to monitor

closely.

● Value: Aquaculture data usually bears a high

volume of information generated by every

stage, [43]. It also offers a lot of potential and

worth for judgments in the future, [44].

Sensors, APIs, post-production research, and

environmental elements are frequently used as

data sources, [45]. It is extremely valuable at

various phases of fish production for

information on water pH, nutrient content, feed

content, humidity, required temperature,

illnesses, and other important details, [46].

The claims mentioned above make it

abundantly evident that fish farming data should be

deemed Big Data in all respects and qualities,

necessitating the adoption of the Data Lake

architecture. The usage of fish farming Big Data

by agriculturists, fish farmers, researchers,

academicians, and decision-makers will determine

its potential.

6 Functional Architecture Proposal

based on the Data Lake

The functional architecture based on a Data Lake

has evolved from mono-zone to multi-zone,

depending on the technical need.

Fig. 3: Functional architecture proposal for Data-driven fish farming system

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

316

Volume 19, 2022

The Data Lake architecture is designed to

support the storage of all structured, semi-

structured, and unstructured data flowing from the

sources. As a result, the mono-zone Data Lake

stores data in its raw format efficiently.

However, in most circumstances, we may

make specific changes to the stored data at the

Data Lake level. Under these conditions, we began

utilizing the multi-zone Data Lakes by establishing

multiple data storage levels. This strategy will

permit us to manage data more effectively. As

illustrated in Figure 3, our suggested Data Lake

architecture for the fish farming context consists of

three data storage levels.

6.1 Data Sources

There might be numerous data sources in fish

farming systems, each with its unique data format.

Our system's primary data source is IoT devices

that capture streaming data of water parameters

such as Pressure, Temperature, pH, Humidity,

Dissolved Oxygen, or salinity. The following data

source consists of data generated manually by

company personnel as CSV or text files. This data

source includes marketing data and measurement

data that sensors still need to automate. APIs

provide data on the weather, the global average

price of fish, and statistics about fish farming

marketplaces.

6.2 Data Lake

The Data Lake architecture features three levels of

data storage. The raw zone is the first level of data

storage, which includes source data without any

alterations. The data can be ingested in real-time or

in batches. This Data Lake Zone offers the data

engineers the ability to find the original data

version.

In the Refined Zone, data can be transformed

based on requirements, and intermediate data can

be stored. In the Refined Zone, there are two data

processing types: batch-based and stream-based.

The Access Zone is the level where data can be

explored. This level enables self-service data

consumption for Reporting, Business Intelligence,

machine learning algorithms, and statistical

analysis.

The Data Governance Zone encompasses all

previous Data Lake Zones to assure data quality,

metadata management, and data accessibility.

6.3 Data Consumption

The data can be consumed in various ways,

including Data Visualization via dashboards

displaying KPIs (Key Performance Indicators) for

each use case, predictive analysis via Machine

Learning algorithms, and statistical analysis.

7 Technical Architecture Proposal

using Big Data Technologies

In order to demonstrate the necessity for Big Data

in a fish farming system, we offer a technical

architecture capable of managing the data flows

generated by normal fish farming activities. In this

architecture, there are three distinct phases (Figure

4):

Fig. 4: Data Management Process with Detailed

Technologies

7.1 Data Acquisition

Technical architecture is a specific set of rules and

interactions of the system's components or features

that ensure the system meets a defined aim and a

set of requirements. For this reason, we suggested

a technical architecture based on Big Data

technologies that enable intelligent fish farming

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

317

Volume 19, 2022

systems.

The design of fish farming creates three sorts

of data. Sensors generate a vast amount of data,

which must be processed before usage. Since there

are numerous devices, it is necessary to translate or

standardize data into a single format. The sensors

communicate with the cloud to transmit data using

HTTP/HTTPS or MQTT protocols. The data

focuses on temperature, light, chemical

composition, and average fish weight, [46].

Complementary data is hand-written in CSV

format and provided through email in flat files.

The files are received and transmitted to the Data

Lake cluster by a data bot, which is a type of mail

bot. Manual data may include the amount of food

per tank or the number of fish. All external data on

the weather, fish prices, fish food prices, and

others are contained within API data. It is retrieved

from APIs using shell scripts and stored in flat files

in preparation for consumption.

7.2 Data Ingestion / Processing

Before data storage, analysis, and access, data

intake collects and delivers data to the processing

system. Based on the type of data, two tools can be

distinguished:

- Apache Oozie is utilized for process

orchestration. It is used to construct and schedule

data jobs at specific dates, times, or frequencies.

Long-running jobs are used to store flat files from

emails and APIs in the Hadoop cluster, [45]. These

activities are scheduled regularly to ensure that

dashboards and reports utilize the most recent data.

- Apache Kafka is responsible for reading and

retrieving the sensor-generated and publishing

streaming data to the OPC server, [45].

Hive tables containing batch data from APIs

and manually fetched data are uploaded to the Data

Lake via Oozie. Apache Spark reads, transforms,

and stores data in additional Hive or Apache

HBase tables.

7.3 Data Consumption

After the data have been prepared comes the data

exposition step. Depending on the use case, data is

presented through dashboards using data

visualization tools, or we use machine learning

algorithms and statistical analysis to explore and

uncover hidden patterns and correlations, [47],

[48]. All this is to allow businesspeople, namely

fish farmers and researchers, easy access to

valuable information, [49].

8 Related Works

Existing research articles are primarily concerned

with agriculture and extracting useful information

from collected data. In their research, Lytos et al.

describe agriculture as a complex scientific topic

that necessitates a particular infrastructure for

managing and interpreting incoming data.

In addition, they gave a comparison of

numerous frameworks used to manage agricultural

data. However, only two of the fourteen

frameworks enable Big Data. Indeed, a specific

Big Data architecture is essential because it

enables the incorporation of IoT, enabling efficient

data collection via sensors and automation of

several tasks, [9].

In parallel, N.N.Misra et al. showed in their

research that since implementing IoT solutions in

agriculture, it has been creating vast volumes of

data, classified as "Big Data," which may provide

new options for monitoring agriculture and food

operations. They discussed how Big Data, IoT, and

AI could be combined to determine the future of

agri-food systems. This study focuses primarily on

agricultural analyses. It focuses on how AI can

monitor greenhouses and how Supply Chain

modernization can improve food quality, [50].

However, they should have discussed how Big

Data approaches contribute to the development of

the fish farming area through IoT integration.

In addition, Xinting Yang et al. highlighted in

their research study how Deep Learning (DL)

might be applied to smart fish farming to address

various data processing difficulties. In addition, he

asserted that the most important contribution of DL

is its capacity to extract characteristics

automatically, [22]. Moreover, before we begin

applying DL to this data, we must be able to

control it, especially when discussing the IoT-

generated enormous data from fish farming.

Creating a specialized Data Lake architecture is the

best way to overcome this scenario so that data

may be more efficiently managed and consumed.

9 Conclusion and Future Work

These days, Big Data methods are used in

practically every business, from finance and

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

318

Volume 19, 2022

banking to manufacturing and advertising to even

medical. In addition, its efficacy has been

demonstrated by its contribution to enhancing

conventional operational procedures and increasing

revenue. Incorporating this technology into fish

farming is a crucial aspect of boosting fish farming

production to meet the expanding global

population's high food demand. However, Big Data

is not frequently used in agriculture or aquaculture

in particular. Big Data tools are the cornerstone of

transforming conventional fish farming into

modern, intelligent, digital fish farming. It solves

the restrictions associated with fish farming

systems by examining farmers' demands, market

needs, financial efficiency, and the viewpoints of

other stakeholders.

This study focuses on fish farming in Morocco

since it reveals a substantial disparity between fish

output and market demand, which results in a

substantial quantity of fish imports. This paper

emphasizes the need for Big Data integration in

aquaculture and subsequently proposes a Data

Lake architecture tailored to the aquaculture use

case.

We propose a functional architecture of fish

farming systems consisting of three steps,

beginning with data sources, moving through the

Data Lake solution, and concluding with data

consumption to ensure the efficient utilization of

generated fish farming data. The data source level

consists of sensor-generated streaming data, flat

files providing additional operational data, and API

data. The Data Lake stage includes a raw zone, a

refined zone, an access zone, and data governance

for data accessibility, usability, integrity, and

security. The final layer, data consumption, is

responsible for data analysis and visualization.

The previously proposed functional

architecture is extended and used to propose a

technical architecture. The proposed technical

architecture relies on three phases: data

acquisition, processing, and consumption. We

detailed each phase and specified its different

technologies and tools, guaranteeing efficient data

handling.

This work's contribution resides mainly in

proposing a functional architecture to exploit the

data generated naturally by fish farms to generate

value through increased productivity and decreased

waste and fish mortality. Then, by proposing a

technical architecture on top of the functional

architecture, we show the feasibility of this

proposal using specific technologies.

Following the proposal of the technical

architecture of the data-driven fish farming system,

our future work will focus on utilizing the

proposed Data Lake architecture as a foundation

for an advanced study employing various forms of

data analysis, such as artificial intelligence and

machine learning.

References:

[1] Bradley, D., Merrifield, M., Miller, K. M.,

Lomonico, S., Wilson, J. R., & Gleason, M.

G. (2019). Opportunities to improve fisheries

management through innovative technology

and advanced Data systems. Fish and

fisheries, 20(3), 564-583.

https://doi.org/10.1111/faf.12361.

[2] Bernhardt, H., Bozkurt, M., Brunsch, R.,

Colangelo, E., Herrmann, A., Horstmann, J.,

... & Westerkamp, C. (2021). Challenges for

agriculture through industry 4.0. Agronomy,

11(10), 1935.

https://doi.org/10.3390/agronomy11101935.

[3] Sarker, M. N. I., Wu, M., Chanthamith, B.,

Yusufzada, S., Li, D., & Zhang, J. (2019a).

Big Data-Driven Smart Agriculture: Pathway

for Sustainable Development. In 2019 2nd

International Conference on Artificial

Intelligence and Big Data (ICAIBD) (pp. 60-

65). IEEE.

[4] Lynch, A. J., Cooke, S. J., Deines, A. M.,

Bower, S. D., Bunnell, D. B., Cowx, I. G., ...

& Beard Jr, T. D. (2016). The social,

economic, and environmental importance of

inland fish and fisheries. Environmental

Reviews, 24(2), 115-121.

https://doi.org/10.1139/er-2015-0064.

[5] The World Bank. (2016). Aquaculture

Production (Metric Tons) | Data.

www.Data.worldbank.org/indicator/ER.FSH.

AQUA.MT?start=2012. Accessed 26 Mar.

2021.

[6] Adeleke, B., Robertson-Andersson, D.,

Moodley, G., & Taylor, S. (2020).

Aquaculture in Africa: A comparative

review of Egypt, Nigeria, and Uganda vis-a-

vis South Africa. Reviews in Fisheries

Science & Aquaculture, 29(2), 167-197.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

319

Volume 19, 2022

https://doi.org/10.1080/23308249.2020.1795

615.

[7] Liu, S. (2020, October 7). Big Data -

Statistics & Facts. Statista.

https://www.statista.com/topics/1464/big-

Data/. Accessed on Mars 20.

[8] Sarker, M. N. I., Islam, M. S., Murmu, H., &

Rozario, E. (2020). Role of big Data on

digital farming. Int J Sci Technol Res, 9(4),

1222-1225.

[9] Lytos, A., Lagkas, T., Sarigiannidis, P.,

Zervakis, M., & Livanos, G. (2020).

Towards smart farming: Systems,

frameworks and exploitation of multiple

sources. Computer Networks, 172, 107147.

https://doi.org/10.1016/j.comnet.2020.10714

[10] Sarker, M. N. I., Islam, M. S., Ali, M. A.,

Islam, M. S., Salam, M. A., & Mahmud, S.

H. (2019b). Promoting digital agriculture

through big Data for sustainable farm

management. International Journal of

Innovation and Applied Studies, 25(4),

1235-1240.

[11] The World Bank. (n.d.). Aquaculture

production (metric tons) - Morocco | data.

[online] Data.worldbank.org. Available at:

https://data.worldbank.org/indicator/ER.FSH

.AQUA.MT?locations=MA. Accessed 11

August 2022.

[12] Holth, M., & Van der Meer, A. (2018).

Aquaculture business opportunities in

Morocco for Dutch entrepreneurs.

https://www.rvo.nl/sites/default/files/2018/0

6/Aquaculture-Business-Opportunities-

Morocco.pdf. Accessed on February 27.

[13] Trade Map. (2020). Fishery products imported

by Morocco.

www.trademap.org/Product_SelCountry_TS.

aspx. Accessed on Mars 14.

[14] Our World in Data. (2018). Fish and seafood

consumption per capita, 1991 to 2017.

https://ourworldinData.org/grapher/fish-and-

seafood-consumption-per-

capita?tab=chart&time=1991..latest&countr

y=~MAR. Accessed on February 27.

[15] Wolfert, S., Ge, L., Verdouw, C., &

Bogaardt, M. J. (2017). Big Data in smart

farming–a review. Agricultural systems, 153,

69-80.

[16] Benjelloun, S., El Aissi, M. E. M., Loukili,

Y., Lakhrissi, Y., Ali, S. E. B., Chougrad,

H., & El Boushaki, A. (2020, October). Big

Data Processing: Batch-based processing and

stream-based processing. In 2020 Fourth

International Conference On Intelligent

Computing in Data Sciences (ICDS) (pp. 1-

6). IEEE.

https://doi.org/10.1109/ICDS50568.2020.92

68684.

[17] Sundmaeker, H. (2016). Accelerating

System Development for the Food Chain: A

Portfolio of over 30 Projects, Aiming at

Impact and Growth. International Journal on

Food System Dynamics, 7(4), 371-381.

[18] Lioutas, E. D., & Charatsari, C. (2020). Big

Data in agriculture: Does the new oil lead to

sustainability? Geoforum, 109, 1-3.

https://doi.org/10.1016/j.geoforum.2019.12.0

19.

[19] Roukh, Amine, et al. "Big Data Processing

Architecture for Smart Farming." Procedia

Computer Science 177 (2020): 78-85.

[20] AMORA, E. N. O., ROMERO, K. V., &

AMOGUIS, R. C. (2020, August).

AQUATECH: A SMART FISH FARMING

AUTOMATION AND MONITORING

APP. In Proceeding of the International

Virtual Conference on Multidisciplinary

Research (IVCMR) (Vol. 27, p. 28).

[21] Bajpai, R., Singh, R., Gehlot, A., Singh, P.,

& Patel, P. (2019, March). Water

Management, Reminding Individual and

Analysis of Water Quality Using IoT and

Big Data Analysis. In International

Conference on Advances in Engineering

Science Management & Technology

(ICAESMT)-2019, Uttaranchal University,

Dehradun, India.

https://dx.doi.org/10.2139/ssrn.3394697.

[22] Mengistu, S. B., Mulder, H. A., Benzie, J.

A., & Komen, H. (2020). A systematic

literature review of the major factors causing

yield gap by affecting growth, feed

conversion ratio and survival in Nile tilapia

(Oreochromis niloticus). Reviews in

Aquaculture, 12(2), 524-541.

https://doi.org/10.1111/raq.12331.

[24] Hu, Z., Li, R., Xia, X., Yu, C., Fan, X., &

Zhao, Y. (2020). A method overview in

smart aquaculture. Environmental

Monitoring and Assessment, 192(8), 1-25.

https://doi.org/10.1007/s10661-020-08409-9.

[25] Wen, Y., Li, M., & Ye, Y. (2020, April).

MapReduce-Based BP Neural Network

Classification of Aquaculture Water Quality.

In 2020 International Conference on

Computer Information and Big Data

Applications (CIBDA) (pp. 132-135). IEEE.

[26] Peng, Z., Chen, Y., Zhang, Z., Qiu, Q., &

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

320

Volume 19, 2022

Han, X. (2020, April). Implementation of

water quality management platform for

aquaculture based on big Data. In 2020

International Conference on Computer

Information and Big Data Applications

(CIBDA) (pp. 70-74). IEEE.

[27] Song, Y., & Zhu, K. (2019, November).

Fishery Internet of Things and Big Data

Industry in China. In 2019 International

Conference on Machine Learning, Big Data

and Business Intelligence (MLBDBI) (pp.

181-185). IEEE.

[28] Rharrhour, H., Wariaghli, F., Goddek, S.,

Sadik, M., & El, A. (2022, June). Towards

sustainable food productions in Morocco:

Aquaponics. In E3S Web of Conferences

(Vol. 337, p. 03004). EDP Sciences.

[28] Yang, X., Zhang, S., Liu, J., Gao, Q., Dong,

S., & Zhou, C. (2021). Deep learning for

smart fish farming: applications,

opportunities and challenges. Reviews in

Aquaculture, 13(1), 66-90.

[29] Parra, L., Sendra, S., García, L., & Lloret, J.

(2018). Design and deployment of low-cost

sensors for monitoring the water quality and

fish behavior in aquaculture tanks during the

feeding process. Sensors, 18(3), 750.

[30] Luna, M., Llorente, I., & Cobo, A. (2019).

Determination of feeding strategies in

aquaculture farms using a multiple-criteria

approach and genetic algorithms. Annals of

Operations Research, 1-26.

https://doi.org/10.1007/s10479-019-03227-

[31] O'Donncha, F., & Purcell, M. Methodologies

for big Data mining in aquaculture.

[32] Kahn, M. G., Mui, J. Y., Ames, M. J.,

Yamsani, A. K., Pozdeyev, N., Rafaels, N.,

& Brooks, I. M. (2022). Migrating a research

data warehouse to a public cloud: challenges

and opportunities. Journal of the American

Medical Informatics Association, 29(4), 592-

600.

[33] Aissi, E., El Mehdi, M., Benjelloun, S.,

Loukili, Y., Lakhrissi, Y., Boushaki, A. E.,

... & Elhaj Ben Ali, S. (2022). Data Lake

Versus Data Warehouse Architecture: A

Comparative Study. In WITS 2020 (pp. 201-

210). Springer, Singapore.

https://doi.org/10.1007/978-981-33-6893-

4_19.

[34] Fleming, A., Jakku, E., Lim-Camacho, L.,

Taylor, B., & Thorburn, P. (2018). Is Big

Data for big farming or for everyone?

Perceptions in the Australian grains industry.

Agronomy for Sustainable Development,

38(3), 1-10. https://doi.org/10.1007/s13593-

018-0501-y.

[35] Kour, V. P., & Arora, S. (2020). Recent

Developments of the Internet of Things in

Agriculture: A Survey. IEEE Access, 8,

129924-129957.

https://doi.org/10.1109/ACCESS.2020.3009

298.

[36] Panwar, A., & Bhatnagar, V. (2020). Data

lake architecture: a new repository for data

engineer. International Journal of

Organizational and Collective Intelligence

(IJOCI), 10(1), 63-75.

[37] Kitchin, R., & McArdle, G. (2016). What

makes Big Data, Big Data? Exploring the

ontological characteristics of 26 datasets. Big

Data & Society, 3(1), 2053951716631130.

[38] Hasan, M. (2020). Real-time and low-cost

IoT-based farming using Raspberry Pi.

Indonesian Journal of Electrical Engineering

and Computer Science, 17(1),

https://doi.org/10.11591/ijeecs.v17.i1.pp197-

204.

[39] Sagar, B. M., & Cauvery, N. K. (2018).

Agriculture data analytics in crop yield

estimation: a critical review. Indonesian

Journal of Electrical Engineering and

Computer Science, 12(3), 1087-1093.

[40] Hajjaji, Y., Boulila, W., Farah, I. R.,

Romdhani, I., & Hussain, A. (2021). Big

Data and IoT-based applications in smart

environments: A systematic review.

Computer Science Review, 39, 100318.

https://doi.org/10.1016/j.cosrev.2020.100318

[41] Lioutas, E. D., Charatsari, C., La Rocca, G.,

& De Rosa, M. (2019). Key questions on the

use of Big Data in farming: An activity

theory approach. NJAS-Wageningen Journal

of Life Sciences, 90, 100297.

https://doi.org/10.1016/j.njas.2019.04.003.

[42] Lee, J., Angani, A., Thalluri, T., & Jae Shin,

K. (2020, January). Realization of Water

Process Control for Smart Fish Farm. In

2020 International Conference on

Electronics, Information, and

Communication (ICEIC) (pp. 1-5). IEEE.

https://doi.org/10.1109/ICEIC49074.2020.90

51285.

[43] Schuster, J. (2017). Big Data ethics and the

digital age of agriculture. Resource

Magazine, 24(1), 20-21.

[44] Carbonell, I. (2016). The ethics of Big Data

in big agriculture. Internet Policy Review,

5(1). https://ssrn.com/abstract=2772247.

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

321

Volume 19, 2022

[45] Pham, X., & Stack, M. (2018). How data

analytics is transforming agriculture.

Business horizons, 61(1), 125-133.

[46] Majumdar, J., Naraseeyappa, S., & Ankalaki,

S. (2017). Analysis of agriculture data using

data mining techniques: application of Big

Data. Journal of Big Data, 4(1), 1-15.

https://doi.org/10.1186/s40537-017-0077-4.

[47] Zhou, C., Xu, D., Chen, L., Zhang, S., Sun,

C., Yang, X., & Wang, Y. (2019).

Evaluation of fish feeding intensity in

aquaculture using a convolutional neural

network and machine vision. Aquaculture,

507, 457-465.

[48] Zhao, J., Li, Y., Zhang, F., Zhu, S., Liu, Y.,

Lu, H., & Ye, Z. (2018). Semi-supervised

learning-based live fish identification in

aquaculture using modified deep

convolutional generative adversarial

networks. Transactions of the ASABE,

61(2), 699-710.

[49] Salman, A., Siddiqui, S. A., Shafait, F.,

Mian, A., Shortis, M. R., Khurshid, K., ... &

Schwanecke, U. (2020). Automatic fish

detection in underwater videos by a deep

neural network-based hybrid motion learning

system. ICES Journal of Marine Science,

77(4), 1295-1307.

[50] Misra, N. N., Dixit, Y., Al-Mallahi, A.,

Bhullar, M. S., Upadhyay, R., &

Martynenko, A. (2020). IoT, Big Data and

artificial intelligence in agriculture and food

industry. IEEE Internet of Things Journal.

https://doi.org/10.1109/JIOT.2020.2998584.

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

Authors Sarah Benjelloun and Mohamed El Mehdi

El Aissi contributed to the research and writing of

the manuscript.

All authors read and approved the final manuscript.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.e

n_US

WSEAS TRANSACTIONS on INFORMATION SCIENCE and APPLICATIONS

DOI: 10.37394/23209.2022.19.33

Sarah Benjelloun, Mohamed El Mehdi El Aissi,

Younes Lakhrissi, Safae El Haj Ben Ali

E-ISSN: 2224-3402

322

Volume 19, 2022