International Journal of Information Management Data Insights 4 (2024) 100210

Contents lists available at ScienceDirect

International Journal of Information

Management Data Insights
journal homepage: www.elsevier.com/locate/jjimei

Critical review on water quality analysis using IoT and machine

learning models
Poornima Jayaraman a, *, Kothalam Krishnan Nagarajan b, Pachaivannan Partheeban c,
Vallidevi Krishnamurthy d
a
Department of Electronics and Communication Engineering, Chennai Institute of Technology, India
b
Department of Electrical and Electronics Engineering, SSN College of Engineering, India
c
Department of Civil Engineering, Chennai Institute of Technology, India
d
Department of Computer Science and Engineering, SSN College of Engineering, India

A R T I C L E I N F O A B S T R A C T

Keywords: Water quality and its management are the most precise concerns confronting humanity globally. This article
Water quality indices evaluates the various sensors used for water quality monitoring and focuses on the water quality index
Internet of Things considering the multiple physical, chemical, and biological parameters. A Review of Internet of Things (IoT)
Sensor technology
research for water quality monitoring and analysis, sensors used for water quality can help remote monitoring of
GIS
and Integrated Technology
the water quality parameters using various IoT-based sensors that convey the assembled estimations utilizing
Low-Power Wide Area Network innovations. Overall, the IoT system was 95 % accurate in measuring pH,
Turbidity, TDS, and Temperature, while the traditional method was only 85 % accurate. Also, this study
reviewed the different A.I. techniques used to assess water quality, including conventional machine learning
techniques, Support Vector Machines, Deep Neural Networks, and K-nearest neighbors. Compared to traditional
methods, machine learning and deep learning can significantly increase the accuracy of measurements of
groundwater quality. However, various variables, such as the caliber of the training data, the water quality
metrics’ complexity, and the monitoring frequency, will affect the accuracy. The geographical information
system (GIS) is used for spatial data analysis and managing water resources. The quality of its data is also
reviewed in the paper. Based on these analyses, the study has forecasted the future sensors, Geospatial Tech­
nology, and machine learning techniques for water quality monitoring and analysis.

Introduction plays a vital role in assessing and monitoring the health and safety of
water sources. It involves evaluating various physical, chemical, and
Water is a highly significant need of every single living being. With biological parameters to determine whether the water quality meets the
an increase in the human population, the requirement for preserving required standards for its intended use, such as drinking purposes, rec­
water assets is acquiring greater significance. Several parameters are reational activities, and ecosystem preservation. Traditional water
used for identifying water quality based on the samples’ water level, pH quality monitoring systems have relied on conventional laboratory-
level, turbidity, and salt content. The machine learning techniques are based methods in which water samples are collected from various re­
combined with a GIS based on data availability and quality. Machine sources and transported to the laboratory for analysis. Water samples are
learning algorithms must account for spatial autocorrelation and tem­ collected manually from different points of interest, such as rivers, lakes,
poral dynamics to predict water quality. Incorporating spatial analysis and wells, and detailed studies are performed using specialized equip­
techniques within GIS, such as geostatistics, can help capture these ment and techniques in laboratory conditions. The obtained data is
variations and improve the accuracy of predictions. By addressing these analyzed and interpreted to assess the water quality. It is compared
problems, robust machine learning and GIS-based frameworks can be against the established standards or guidelines to meet the criteria for its
determined for water quality index calculation, aiding in efficient water designated purpose. The limitations of traditional laboratory-based
management and decision-making processes. Water quality analysis water quality monitoring systems are that they are time-consuming

* Corresponding author.
E-mail address: [email protected] (P. Jayaraman).

https://doi.org/10.1016/j.jjimei.2023.100210

Available online 4 January 2024

2667-0968/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

and costly, lack real-time monitoring, and are inaccessible to remote R2: What strategies can be figured out by associations to develop
locations. robust water quality forecasts using machine learning models that align
with sustainability objectives?
Data-driven water quality assessment The answer to the above questions is that the present study Focuses
on the intersection of IoT, A.I., and GIS in water quality expectations.
Water is a fundamental resource that has always supported different This review surveyed 67 articles from driving diaries in the Scopus data
human activities, including farming, industry, and homegrown use. set (Sharma et al., 2021). This strategic methodology permits us to
However, the nature of water assets is progressively compromised by introduce thorough bits of knowledge and exploratory discoveries.
anthropogenic activities, such as contamination from modern releases, Sections 2 and 3 focus on literature reviews and comparison of various
farming overflow, and urbanization. Checking and evaluating water Review methodologies. Section 4 illustrates the findings of the system­
quality is essential for ensuring the supportability of water assets and atic literature review on GIS, and Section 5 discusses the machine
safeguarding human well-being. Traditional methods of water quality learning algorithms and multiple techniques implemented in various
assessment deal with manual sampling and laboratory analysis. It is a applications in the research area. Section 6 is the conclusion.
time-consuming process. (Uddin et al., 2021) have developed the traditional methods of WQI
The advent of data-driven technologies, including sensors, the models. Due to its generalized structure, the evaluation cites the con­
Internet of Things (IoT), A.I., and Geographic information systems (GIS), version of substantial data on water quality into a single value or index
has changed the field of water quality evaluation. Information-driven as one of the valuable techniques. They divided WQI models into four
approaches offer a few benefits over conventional techniques, consecutive steps: selection of parameters of water sub-indices genera­
including Real-time monitoring, High-resolution data, spatial and tem­ tion, parameters weighting values, and overall WQI calculation.
poral analysis of water quality, and predictive modeling. The data- Therefore, the large volume of water quality data is compressed into a
driven water quality assessment is essential for addressing the single index with a comparison between various traditional WQI Index
following challenges: Increasing water scarcity, pollution from nonpoint related to selecting parameters, sub-index procedure, weighting, ag­
sources, and the impact of climate change. Data-driven water quality gregation techniques, and rating scale. (Nayak et al., 2020) analyzed the
assessments to ensure the availability of clean and safe water for future water quality status of Indian rivers.
generations. The framework of water quality data analysis proposed by (Bui et al., 2020) Assessed WQI and concluded that traditional WQI
WHO and BIS. calculations consumed more time and identified few errors by calcu­
This literature review explores the potential of data-driven ap­ lating sub-indices. Numerous statistical and visual evaluation indicators
proaches in accurately predicting water quality index (WQI) using ma­ were used to evaluate the models. The machine learning algorithm
chine learning and GIS techniques. The study will focus on the following divided the data into training and testing sets, used hybrid algorithms,
key aspects: and estimated WQI Values. (Wu et al., 2017) to obtain WQI analyses, the
research work was carried out in Lake Poyang, China, where 24 samples
1. Machine learning algorithms: Identify and evaluate the most effec­ of water quality values were classified into three groups. This work
tive machine learning algorithms for WQI prediction. considered 20 different water quality parameters for analyzing Total
2. GIS integration: Explore the role of GIS in spatial analysis and visu­ Nitrogen (TN) and Total Phosphorus (TP), whereas hazardous metals
alization of WQI data. and other criteria were rated low in the WQI Analysis. (Wang et al.,
3. Data considerations: Address the challenges and opportunities 2017) Worked with a broad classification of different fractional devia­
associated with data collection, preprocessing, and management for tion methods referred to difference, ratio, and normalized difference
data-driven WQI prediction. index. The obtained WQI values vary from 56.61 to 2886.51. The above
4. Performance evaluation: Discuss the metrics and methodologies for index was estimated using the slope of the curve and the root means
evaluating the performance of data-driven WQI prediction models. square error values.
5. Case studies: Analyze real-world case studies to demonstrate the (Lamare & Singh, 2016) Did the work on Numerus’s pollution index
practical application of data-driven WQI prediction in various water (NPI), the simplest WQI evaluation method known among all the WQI
resource management scenarios. methods. Their study evaluated the specific parameters for their po­
tential contribution to water bodies. (Dohare et al., 2014) Examined
Review of analysis of water quality parameters and indexes water quality for Indore City, and they calculated WQI using around 27
parameters: pH, color, Total Dissolved Solids (TDS), calcium, chromium,
Various physio-chemical and biological parameters for measuring electrical conductivity (E.C.), total alkalinity, hardness, zinc, manga­
and monitoring the Water quality index (WQI) exist worldwide. The nese, and nickel. The evaluated results are as per IS 10500–2012 spec­
assessment procedure of WQI is analyzed and summarized in this sec­ ifications. They used statistical methods to analyze the physio-chemical
tion. Analyses of various water quality measurement parameters have parameters and water pollutants characterization.
been made using traditional methods that help to calibrate the physical (GG.S, 2013) A standard formula for WQI was used by fixing the
changes. (Wu et al., 2021) Carried out experimental work on the critical numeric Q-value and weightage factor. Despite fixing the weight factors
parameters that affect water quality. Fifteen physio-chemical parame­ for several parameters, they adopted the dissolved oxygen level, fecal
ters were evaluated to assess the WQI and categorized into three groups. coliform, B.O.D., pH, nitrate, Phosphorus, temperature, turbidity, and
These helped establish the results with minimum error and good accu­ TDS Using these conditions, they arrived at an equation to calculate WQI
racy. This detection benefits low-cost and fast-track water quality with missing parameters. (Nayar, 2020) Analyzed Physio-Chemical pa­
evaluation in adjacent basins. Based on the statistical values, the rameters: temperature, Transparency, hardness, pH, sulfate, chloride, D.
spatially evaluated distribution pattern and different models of WQI are O., B.O.D., C.O.D., alkalinity nitrates phosphates, and water quality.
classified. Given the expansive implications of water quality prediction They classified parameters as (0–25) - Excellent quality, (26–50) - Good
utilizing a complex methodology, our review aims to address funda­ quality, (51–75) - poor quality, (76–100) Inferior quality, >100 un­
mental research questions, making significant contributions to the field. suitable for drinking water.
The recognized research questions directing our examination are as per Finally, (Uddin et al., 2021) employed the Analytic Hierarchy Pro­
the following (Kar & Varsha, 2023): cess (A.H.P.) to categorize the World Quality Index as exceptional
R1: How do integrating Internet of Things (IoT), AI, and Geographic (90–100), good (90–75), fair (75–50), marginal (50–25), and low
Data system (GIS) functionalities add to the headway of water quality (25–5). This proves to be a very efficient method for minimizing un­
expectations? certainty, weighting procedures, and improving accuracy. Table 1

2
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Table 1
Comparison of WQI Models weighted parameters used by various researchers.
Parameters Horton ( National Sanitation W.J.WQI ( Environmental quality Scottish Research Bascarron Index Said ( Oregon (
Jha et al., Foundation ( Uddin et al., index (Uddin et al., Development Department (BWQI) (Sun Uddin Swamee &
2015) Dawood et al., 2022) 2021; Liu 2021) Index (Sutadian et al., et al., 2016) et al., Tyagi, 2000)
et al., 2019) 2016) 2021)

pH 4 0.11 – 0.1 0.09 1 – 0.16

Turbidity 0.08 – 0.1 – 4 0.1 –
D.O. 4 0.17 0.10 0.1 0.18 4 1.5 0.16
NO3 0.10 0.06 0.15 – 2 – 0.16
Cl 1 – 0.07 0.1 – 1 – –
Temp – 0.10 0.03 0.1 0.05 1 –
T.S. – 0.07 – 0.1 – – – –
B.O.D. – 0.11 – 0.1 0.15 3 – 0.16
Colour – – 0.1 – 2 –
F.C. 1 0.16 0.17 0.1 0.12 – 15 0.16

presents details of the different kinds of WQI models and weightage used five contributing variables: lithology, drainage, land cover/land use,
by various researchers. Several authors have analyzed some parameters. lineaments, seepage, and slope. The outcomes showed that around 1.2%
A comparative analysis by different authors shows that pH analysis of of the review region was assigned as having great potential for
the Horton index shows better results than any other index method. The groundwater to re-energize (Yeh et al., 2016). The curve between the
Bascarron index shows better results for calculating turbidity, and other wavelength and reflectance of the WQI value for different sample sites is
remaining parameters showed better outcomes for Bascarron discussed. Analytical hierarchy-based images are used to identify suit­
index-based weightage results. A cluster analysis of water quality spatial able solid waste dumps based on the assigned weights of the various
distribution given by various research studies is presented in Table 2. parameters. The map was prepared to classify the trustworthy dump site
as highly qualified, less fortunate, and unsuitable for the site selection.
Review methodology So, the multi-criteria decision-making analysis was done to identify the
location, presented in Fig. 2.
There are several proposals for a writing survey. It involves a four- Researchers and water resource professionals effectively integrate an
step approach for doing the writing survey, as depicted in the ensuing GIS and visualize the various groundwater data layers by employing a
parts. This technique helps us recognize and concentrate on topics and methodological approach for groundwater analysis using GIS This
future exploration headings. Fig. 1 shows the flowchart of the review approach enables a better understanding of the spatial patterns, trends,
methodology adopted for Water quality prediction using Sensors, IoT, A. and relationships between groundwater parameters and environmental
I., and GIS technologies. factors. The integration of GIS with hydrological modeling and decision
support systems facilitates informed decision-making, sustainable
groundwater management, and protection of valuable resources. The
Review of water quality analysis using GIS
methodology describes the data acquisition system using wired and
wireless sensors. Hydrological modeling tools simulate the groundwater
The WQI can be predicted by plotting the spatial characteristics of
flow and evaluate the water availability. Various data sets are pre­
water in lakes (Wang et al., 2017). Some of the parameters were esti­
processed, and model selection, training, and evaluation of multiple
mated using remote sensing monitoring Technology. Depending on the
metrics are predicted. Some examples where the contexts based on GIS
problem domain, the boxplot analysis provides an accurate visualization
are discussed for providing ideas about spatial and temporal mapping
prediction and finalizes the outlier detection of threshold values. Ma­
along with the implementation of A.I. models are listed in Table 3.
chine learning techniques were applied based on the captured image
data sets from the Gravity Recovery and Climate Experiments (GRACE).
Review of sensors-based water quality monitoring systems
These helped with accurate water quality predictions and images of
groundwater distribution.
Continuous monitoring of water quality parameters can be measured
(Nampak et al., 2014) A GIS approach was utilized to incorporate

Table 2
Water Quality Parameters Utilized by Different Researchers.
Author Reference

Water parameters (El Bilali & Taleb, (Jha et al., (Dohare et al., (Ewaid et al., (Ewaid et al., (Chen et al., (Shakhari & Banerjee,
2020) 2015) 2014) 2019) 2018) 2020) 2019)

pH 7.90 7.567 7.70 7.909 7.43 – 7.979

Conductivity (μS) – 585.3 – – – – –
Water temperature 20.1 16.13 – 33.00 – – –
Turbidity – 38.35 40.36 63.87 – – –
B.O.D. (mg L− 1) – – 9.381 2.98 2.23 – 7.80
Dissolved oxygen (mg 8.69 – 5.67 – – – 10.4
L− 1 )
Total Nitrogen (mg L− 1) 3.95 – 6.50 – 3.47 – 7.06
Ammonia (mg L− 1) 0.80 – – – – –
Nitrate (mg L− 1) 1.13 – – – – – –
Total phosphorous (mg 0.27 – 0.058 – – – 1.7
L− 1 )
Calcium (mg L− 1) 23.8 – – 224 92.0 – –
Chloride (mg L− 1) 23.7 – – – – – 883.8
Sulphate (mg L− 1) 22.8 – 91.36 – 358.15 – –

3
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Fig. 1. Flowchart on review methodology using various approaches for water quality prediction (Kar & Varsha, 2023).

Fig. 2. Methodological approach for groundwater analysis using GIS (Arabameri et al., 2021).

using different portable and low-cost approaches. (Tuna et al., 2013) Prediction & Simulation Package). The lifetime of node and trans­
Focused their work on the practical usage of mini boat probes loaded mission power was analyzed, along with variations in the transmission
along with the wireless sensor network monitoring system (WW.S.N.. power of the nodes between 0 dB and − 25 dB. Fig. 3 shows the wireless
IEEE 802.15.4 is a wireless interface system meant for continuous water sensor network-based monitoring system. A field test in the water
quality monitoring, and a few parameters, including DD.O. pH, EE.C. reservoir was done in Kirklareli Dam. A web-based online interface with
temperature, turbidity, and nitrate, to determine water quality, and remote monitoring IoT devices controlled the system. Researchers were
simulation studies were focused on monitoring the water quality of focused on the parameters such as D.O. pH, EE.C. temperature,
drinking water reservoirs. (Wu et al., 2021) analyzed the maximum turbidity, and nitrate. They concluded the simulation results using a
water quality parameters meant for better accuracy. (EE.L.Watershed web-based interface and the system’s dependence on parameters such as
et al., 2017) Analyzed a few parameters to get the nearest accurate transmission frequency, power, size of the packet node parameters, etc.
values. It is important to note that the sensors monitoring the water quality
The system integrated with GG.P.S. provides an online data acqui­ should be cost-effective and continuous. The micro sensors fabricated
sition system with two long-range radio modems for simultaneous data along with the channel interface should measure water quality param­
transmission. The MATLAB-based simulation environment was also used eters such as pH and conductivity. The final results were compared with
for analyzing the water quality parameters with the source codes of MM. the commercial sensors in static and dynamic conditions. Fig. 3 shows
A.T.S.N.L.(MM.A.T.L.A.B. Wireless Sensor Node Platform Lifetime the resistance of the pH sensor measured in static conditions. The same

4
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Table 3 through the deployment of L.P.W.A.N. scan assist in identifying and

survey study on GIS and fitness in water quality assessments (Venkatachalam & responding in polluted situations. Overall, the research on LL.P.W.A.N.s
Ray, 2022). for monitoring the quality of subsurface water still encourages the
Year Authors Review Context discussed application of LL.P.W.A.N.s has a variety of benefits, and they are being
2022 (Dawood et al., Application of Water This study assessed and
applied more frequently.
2022) Quality Index and mapped groundwater (Nayar, 2020) Used wireless sensor network technology with the
Multivariate Statistical quality and spatial GSM Module and ZigBee protocol to create a low-cost water quality
Techniques to Assess and distribution. Multivariate monitoring system. The microcontroller of the electrochemical sensor
Predict Groundwater statistical techniques
was used to create a hardware system with wireless communication
Quality with the Aid of were used to identify
Geographic Information sources of pollution and networks for monitoring water temperature, dissolved oxygen, and pH
System develop a model to values in a pre-programmed period. (Bell, 2013) Authors have employed
predict water quality a wide range of sensors to measure pH, turbidity, water level, temper­
index. ature, and humidity of water. They used a pH sensor, Turbidity sensor,
2019 (Thinsungnoen, An Integrated System of This study developed a
2019) Geographic Information decision tree model to
Ultrasonic sensor, DHT-11 sensor, and microcontroller Node MCU with
and Water Quality: predict water quality E.S.P.Wi-Fi module, think to speak IoT data collection application.
Lamtakong River based on land-use factors (Sanya et al., 2022; Bates et al., 2021). This study uses LoRaWAN to
and created a GIS tool for monitor the water quality in rivers and lakes. The authors emphasize
retrieving and displaying
LoRaWAN’s benefits in terms of long-distance connectivity, low power
water quality data
2018 (Tiwari et al., Understanding the spatial Analyzing the plume’s consumption, and affordability. They discuss the installation of LoR­
2018) and temporal dependence movement serves as a aWAN sensors to gauge variables in water bodies, including pH, tem­
of the migration of proactive step, laying the perature, and dissolved oxygen. The study proves the viability and
conservative contaminant groundwork to effectively dependability of LoRaWAN for monitoring water quality in real-time
plumes in the urban mitigate potential
applications, and the authors discussed the choices for data trans­
groundwater contamination in the
environment in the areas surrounding the mission, network architecture, and sensor selection. In terms of data
Panchkula region, disposal site. accuracy, transmission range, and power consumption, they assess the
Haryana, India system’s performance, and the outcomes show how the LoRaWAN-based
2006 (Ghose et al., GGISbased transportation A proposed GIS optimal
system is accurate and successful in monitoring the water quality met­
2006) model for solid waste routing model aims to
disposal – A case study on identify the most cost- rics. Fig. 4 Shows the graph for the sensor ranges adopted by the
Asansol municipality effective and efficient different researchers to monitor pH values.
collection routes for (Almetwally et al., 2020) designed and developed a system for
transporting solid wastes automatic decision-making to control water quality parameters. The
to landfills.
different water quality sensors resulted in acceptable limits and recor­
ded historical readings in a cloud platform. The system is related to
public health and eliminates the inconvenience faced during offline lab
analysis of collected samples. (Kothari et al., 2021) Authors have used a
variety of sensors, including temperature, TDS pH, and DD.O. sensors to
measure the WQI parameters. The information gathered from the GSM
SIM 900 module and Arduino mega2560 microcontroller communicated
the information to the data set server at Things Peak. The information
sent was seen utilizing an Android IoT application associated with the
server and the. The information showed the effective continuous
application of getting the clients ready to check the sensors on an
Android versatile screen. (Roseela et al., 2020) Implemented submerged
remote correspondence accomplished through IoT made a trend-setting
Fig. 3. W.S.N. based on water quality monitoring sensors (Bell, 2013). innovation with the microcontroller and Wi-Fi. In this review, temper­
ature, turbidity, and distance between the articles were estimated using
sensor is mounted with different pH liquids can find a change in the pH respective sensors associated with the transmitter and recorded at the
value regardless of the exposure time. IoT-based water quality recipient side, utilizing IoT innovation for diverse time stretches
measuring system using sensor-based circuitry powered by solar (Akbarighatar et al., 2023), [57]. The proposed U.W.C. framework was
renewable energy analyzed pH level, turbidity, salinity, arsenic level, seen as dependent on IoT and known for its minimal expense and effi­
etc. The information collected from the sensors was collected in cloud ciency, and it can be utilized for various applications. The sensors used
computing using a Wi-Fi module, and the graphical comprehension
demonstrated data via Think Speak.
(Azpilicueta et al., 2022) Applications for monitoring subsurface
water quality have grown in popularity in recent years in low-power
wide area networks (L.P.W.A.N.s). For this application, L.P.W.A.N.
provides several benefits, including Long-distance communication for
monitoring groundwater quality. L.P.W.A.N. devices offer low power
consumption and help monitor remote applications where replacing the
batteries regularly and running for extended periods is impractical. L.P.
W.A.N.s is cost-effective for monitoring the quality of subsurface water,
and several research works have investigated the use of L.P.W.A.N.s for
monitoring the quality of underground water. A variety of water re­
sources can be monitored using L.P.W.A.N.s and Numerous aspects of
water quality, such as pH, temperature, turbidity, and conductivity, can
be tracked using L.P.W.A.N.s Real-time water quality monitoring Fig. 4. pH Resistance under static conditions.

5
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

for various applications for measuring water quality play an essential to analyse the water quality metrics to make predictions. The final re­
role. Table 4 details various authors’ sensors for measuring water quality sults demonstrated a good indication of the model’s anticipated and
and suitable hardware, software, and networking systems locations to actual values using multidimensional input data sets and assessment for
ensure efficient and accurate output. the future emerging water quality trend using LSTM deep neural net­
works to forecast drinking water quality.
Water quality analysis using machine learning techniques (Shakhari & Banerjee, 2019) The authors used a multi-class classi­
fication system with continuous data sets to regulate water quality. They
Employing IoT (Internet of Things) and machine learning for water efficiently analyzed data using an algorithm known as the decision tree
quality monitoring offers advanced capabilities for collecting, classification model to forecast the water quality. The research is
analyzing, and interpreting data. (Wang et al., 2017) proposed a ma­ employed in water quality monitoring equipment to classify the quality
chine learning technique employing WQI and fractional derivative of bottled drinking water, pipeline drinking water, and pond water taken
methods to get remote sensing spectral indices. A support vector at center stage. They compared the quality of two methods of classifi­
Regression model was implemented to analyse the model’s accuracy. cation and regression. (Sharma et al., 2021) A supervised algorithm and
(Liu et al., 2019) Work on smart water quality analysis and the predic­ machine learning approach accomplished the water quality prediction.
tion was done using Long short-term memory (LSTM) time series neural This approach needed to be updated with drawbacks, such as lab time
network IoT Environment. They used training and test monitoring data consumption, cost, and statistical analysis. Analysis of four different
parameters, namely turbidity, temperature, pH, and TDS, shows that
polynomial regression with a level of 2, angle assistance, and learning
Table 4 rate of 0.1 outperformed other relapse computations by its proficiency in
Review of Hardware, sensor, and I.O.T. techniques used by the researchers. most accurate WQI anticipation.
Sl. Author Sensors and Parameter measured Accuracy (Lamare & Singh, 2016) did the prediction of water quality for irri­
No Reference Hardware tools level gation purposes using six models of machine learning techniques
used
exhibited A.I. models as proficient methods for the precise view of the
1 (Kar & Smart water pH, D.O. conductivity, The mean nature of water systems. Robotized sensor implementation and machine
Varsha, quality sensor temperature, value of
learning models help researchers regulate water quality and address
2023) oxidation–reduction 69.01
potential, turbidity, Accuracy irrigation water quality issues. An automated sensor technology coupled
dissolved ions level:93.08 with K-NN and S.V.R. machine learning provides accurate results.
% (Ewaid et al., 2018) performed groundwater quality prediction using
2 (Wang Optical sensors, pH, D.O. Turbidity, Accuracy machine learning techniques, and the initial step involved in the
et al., 2017) Microelectronics Escherichia Coli, and level:87 %
application of regression approach to predict groundwater images and
Mechanical Residual Chlorine.
systems, and Bio- the process adopted are shown in Fig. 5. Gaussian Mixture Model was
Sensors used to produce minor errors, and the machine learning models were
3 ( Kapta 3000 AC4 Chlorine, – applied in the GRACE Dataset to predict future datasets in the form of
Azpilicueta sensor Spectro: Temperature,
pixels.
et al., 2022) laser pressure,
conductivity,
Data mining algorithms with the combination of standalone and
Turbidity, and color,
4 (Najah The system Metal toxic pollutants Accuracy
Ahmed combines Hg2+, Cu2+ and level:93.3
et al., 2019) computer image Zn2+ %
processing
technology with
biological sensors.
5 (Nayar, DD.O.and pH D.O. pH values Accuracy
2020) sensors and level:96.3
zhorrible water %
checker, WW.S.N.
GSM Technology
6 (EE.L. SCADA, internet in Collection, storage, SCADA
Watershed Raspberry Pi analysis, and Accuracy-
et al., 2017) visualization 94 %
7 (Tuna et al., Temperature pH, turbidity, –
2013) sensor, pH, temperature
Turbidity,
Ultrasonic sensor,
Arduino Uno,
water flow sensor,
Filter unit
8 (Demetillo Data logging pH, T TDS
et al., 2019) system with a 32- Data logger
bit AA.R.M.core Accuracy-
microcontroller 97 %
board based on the
Atmel SAM3 × 8E
ARM Cortex-M3
9 (Liu et al., Arduino and Node pH, Turbidity, E.C.
2019) MCU, PP.H. Temperature, Accuracy of
Sensors, Turbidity Dissolved Oxygen, sensors-
sensor, Dallas and E 91.2 %
Temperature
sensor, DD.O.
sensor, salinity
Fig. 5. Flow chart of an implementation of machine learning models (Hussein
sensor
et al., 2020).

6
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

hybrid algorithms models were evaluated. Faecal coliform (FF.C. and bagging (B.A. and Random tree (RR.T. outperformed the other models,
total solids (TT.S. are the two parameters considered and revealed. and they used a boxplot to predict the errors such as R2, RMSE. MAE, and
(Mancuso et al., 2021) used post-processing techniques produced by NSE. (GG.S, 2013) used an algorithm that works with big data analysis
time series as a forecasting method, and this time series with hierar­ with a reduced cost to improve the prediction efficiency. They observed
chical extracts showed the end Neural Network with an accurate fore­ D.T. R.F. and D.C.F. give better predictions than other algorithms.
cast. The Neural Network disaggregation model was created, tested, and (Venkatachalam & Ray, 2022) The trained requirement is necessary to
trained by different data sets. (Dilmi & Ladjal, 2021) used the Long Short train the algorithm to be more accurate. (Ahmed et al., 2019) They
Term Memory deep learning approach to classify the water quality analyzed the data using a supervised algorithm and considered the error
network of recurrent neurons and achieved an accuracy of 99.72% using and deviations for predicting WQI and WQC He concluded the gradient
models for time series prediction, with popularity in deep learning boosting algorithm is more efficient than other regression algorithms in
techniques. They used a multisensory monitoring system for the con­ estimating MAE. MSE. RMSE and R2. and the polynomial regression
struction of the database. The performance assessment was done using shows better results compared to others. In the present study, the au­
AI techniques, LSTM and RNNs, and multi-class SVM. Different types of thors mainly focus on the WQI and WQC. parameters to predict the
analysis were out for data reduction. The technology used principal future performance and efficiency of different algorithms. This research
component analysis (P.C.A., linear discriminant analysis (LDA), and primarily focuses on comprehending the challenges related to the uti­
independent component analysis (II.C.A..) lization and consequences of the emerging generation of AI-based sys­
(Chen et al., 2020) Did a comparative analysis of water quality tems in decision-making processes. This work’s objective involved
prediction using machine learning with extensive data analysis?. The identifying the potential research opportunities specifically for Infor­
comparison was made between 10 learning models for assessing the mation Systems (II.S.. Water quality monitoring can be done in real-time
quality performance using big data. (Herath & Mittal, 2022) (Yang et al., using data analytics. This can aid in the early detection of potential is­
2019) The data sets collected are related to the parameters pH, D.O. C.O. sues so that they can be fixed before doing significant harm (Pasika &
D. Mn, and NH3 and validated by the model’s Decision tree (D.T., Gandla, 2020). For instance, data analytics can be used to monitor the
Random Forest(RR.F. and Deep cascade Forest(D.C.F.. The best models concentrations of contaminants like microorganisms, heavy metals, and
were identified based on the increased efficiency in predicting surface chemicals in water. These details can be used to pinpoint regions where
water quality. (Aldhyani et al., 2020) Carried out work on water quality there is a chance of poor water quality and to take preventative
prediction using AI Algorithms that includes nonlinear autoregressive measures.
neural network (N.A.R.N.E.T. and long short-term memory (L.S.T.M. (Herath & Mittal, 2022) Analyzed high-impact impacts AI algorithms
deep learning algorithm. Seven datasets using the L.S.T.M. algorithm are used for various smart city domains, and it analysed the future work
shown in Fig. 6. Table 5 analyses the various machine learning algo­ based on collaborating Artificial intelligence with Big data along with
rithms used in water quality analysis by multiple researchers. virtual reality, 3D printing, etc. This work concluded with effective al­
Using an artificial neural network and regression plot, the compar­ gorithms such as ANN, RNN/LSTM, and SVM. Data analytics can be used
ison of SVM, K.N.N. Nave Bayes models were made, and the output from to forecast patterns in water quality in the future. Water quality man­
test data was found to improve the accuracy in the prediction of WW.Q.I agers can use this information to prepare for possible issues and
(Uddin et al., 2021) The authors analyzed the WQI and WQC using implement preventative measures. Data analytics, for instance, can be
machine learning techniques and proved that the GARNET result used to forecast the effects of climate change on the standard of the
demonstration provides a better forecast than LSTM Following that water. (Kalini, 2017) Using this knowledge, measures can be created to
(Dohare et al., 2014) is the prediction of WQI based on hybrid data safeguard water quality in the face of climate change and geographical
mining algorithm. Based on the above techniques, their approach may analysis. Spatial data, such as where water bodies are located and where
consume time and lead to errors and deviations in sub-indices—Stand­ pollutants are distributed, can be analyzed using data analytics. This can
alone (4 algorithms) and hybrid data mining (12 algorithms). Hybrid assist in locating regions most vulnerable to poor water quality. Data
analytics, for instance, can be used to visualize how pollutants are
distributed throughout a watershed. The places that require pollution
control measures the most can be determined using this information.
Data analytics can be used to create machine learning models that can
find trends in data on water quality. These models can be used to fore­
cast future water quality trends and spot potential issues. (Sn et al.,
2020) Machine learning algorithms, for instance, can be used to find
trends in water quality data that are connected to certain pollutants.
Using this data, early warning systems can be created to notify managers
of water quality and potential issues. These are just a handful of the
numerous ways that data analytics can support water quality moni­
toring. Water quality managers can better identify trends and patterns in
water quality by employing data analytics. This information can be
employed to control water quality more effectively.
(Yang et al., 2019) Based on the support vector machine frequency
data, it can be forecasted in short-term high-frequency data. (Duan et al.,
2019) reviews the various advancements in AI algorithms to improve
their computation of storage and power. (Kalini, 2017) (Sn et al., 2020)
I.J.I.M.’s paper used the AI research tool to analyze the various data
mining methods.
Employing IoT and machine learning for water quality monitoring
offers additional capabilities for collecting, analyzing, and interpreting
data using data analytics.

a) IoT for Data collection: IoT devices like sensors and probes are
Fig. 6. Flow diagram for estimating efficient algorithm (Uddin et al., 2021). deployed in water bodies. The collected data is transmitted

7
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Table 5
Analysis of Machine Learning Algorithms implemented for assessing water quality.
Sl. Pilot scale experiments Water quality parameters Machine learning techniques Range Achieved Preferred Reference
No. Proposed algorithm

1 1. Reclaimed water pH, Cl, Conductivity, C.O.D.1. Direct SVM SVM: (DA 97.22 % T.P.R.values PCC-SVM based (Xu et al.,
2. Portable water Temperature, Turbidity, C 2. PCC-SVM-based method 91.67 % FF.P.R.being 0) PCC- method 2019)
3. Single parameter method SVM: (DA> 97 %, TT.P.R. 93 %
and FF.P.R.= 0)
2 Water quality monitoring Water Temperature, pH, 1. L L.S.T.M.2. AA.R.I.M.A.3. SS.V. L.S.T.M. (Limb et. al,
station of Guazhou Water Dissolved Oxygen, C.O.D. R.Period length m = 10 2011)
Source of the Yangtze Conductivity, Turbidity, LSTM-0.0017
River NH3-N ARIMA-0.0126
SVR-0.0046
3 packaged drinking water, pH T.S.Reference Indicator based, MAE-0.17 Reference (Shakhari &
Pipeline drinking water, T Decision Tree, C4.5, .Logistic RME-0.20 Indicator-based Banerjee,
and pond water Regression Decision Tree 2019)
4 Evaluated quality of the pH, BB.O.D. NH+ 3 , Iron, WQImethod, DI, RI and NN.D.I. R2–0.92, RMSE-58.4, RPD-2.81 Particle swarm (Wang et al.,
Ebinur Lake Watershed Copper, Zinc, D.O. TTDS correlation coefficients, SVM model optimisation 2017)
surface water Cl− , Na, etc. (20 (PSO)-support
parameters) vector regression
model
5 A case study Tilesdit dam pH, EC, T, TU Mg, B, H 1. L.S.T.M.RNNs PCA 95.17 %. LDA LSTM RNN (Dilmi &
in Bouira (Algeria) 2. Multi-class SVMs 97.84 %, ICA 95.42 % Ladjal, 2021)
6 Rawal watershed, situated Temperature, turbidity, TDSsupervised prediction MAE =1.96, F1score=0.56 Regression (Ahmed
in Pakistan pH, T algorithms Recall=0.564 Algorithms et al., 2019)
7 GRACE ΔTWS image Full image prediction to Extreme gradient boosting, MAESVR Mean absolute error- support vector (Hussein
reconstruction determine groundwater multivariate linear regression, 2.70, RMSE SVR regression et al., 2020)
quality prediction random forests, multilayer Root Mean Square Error-5.72
perceptron, and support vector
regression
8 Four monitoring stations Ammoniacal Nitrogen, Adaptive Neuro-Fuzzy Inference pH, No. of neuron-18, RMSE- Neuro-Fuzzy (Tiwari et al.,
of Johor River. suspended solid (SS.S., pH System, Radial Basis Function 0.15 Maxerror%− 3.22, R2=0.9 Inference System 2018)
Neural Networks (WDT-ANFIS)
9 Monthly data from 2 water B.O.D. C.O.D. D.O. pH, Random forest, Random tree, R2 = 0.941, RMSE = 2.71, MAE BA-RT (Dohare
quality monitoring turbidity, electrical Reduced error pruning tree, = 1.87, NSE = 0.941, PBIAS= et al., 2014)
stations within the Talar conductivity, and Total Around 12 hybrid data-mining 0.500
catchment solids algorithms
10 Kaggle-Indian water D.O. pH SVM, K-NN, Naive Bayes nonlinear N.A.R.N.E.T. LL.S.T.M. SVM model (Uddin et al.,
quality data sets Conductivity, BB.O.D. autoregressive neural network N Accuracy-97.01 % Sensitivity- (WW.Q.C. 2021)
Nitrate, Fecal coliform, 99.23 % Specificity-97.78% NARNET model
and total coliform Precision-94.93 % (WQI)
R.N.A.R.N.E.T.96:17 %
R.L.S.T.M. 4:21 %

wirelessly to a central database or cloud platform, enabling remote contaminations in most encompassing water bodies like waterways,
monitoring and access to information. lakes, and streams, which have explicit quality guidelines. e. The decay
b) Data Integration and Management: The data management system of water quality impacts oceanic biological systems by reducing the
organizes, stores, and preprocesses the collected data. supply of safe new water for domestic use. Developing nations
c) Data Analysis using Machine Learning: Various algorithms in iden­ frequently go through quick financial expansion, and each improvement
tifying the abnormal water quality or potential pollution sources. It project has the possibility of producing adverse consequences on the
also identifies patterns of water quality data and develops models for climate. Recently, the abundant usage of manure has generally moved
predicting future water quality conditions (Almetwally et al., 2020). into groundwater streams. Moreover, the standards and specifications
d) Prediction models: By training the models on historical data, the vary for different applications/uses. For instance, the water system is
system can identify real-time notifications. neither saline nor contains harmful materials, and it can be moved to
e) Spatial models and statistical analysis: Data analytics are used to plants or soil and accordingly annihilate the environment. In wireless
identify the most affected areas in water quality, which is highly risk. sensor applications, Sondes furnished with sunlight-based chargers for
Statistical analysis tends to identify the patterns in water quality energy harvesting is updated in recent works. Sondes are outfitted with
areas. IEEE 802.15.4-based remote connection points and mounted on floats.
They structure a Remote Sensor Organization to speak with one another
By combining IoT and machine learning, water quality monitoring and send their estimations at standard stretches to a focal PC. through
systems can provide real-time insights, predictive capabilities, and WSN in terms of GIS implementation, the movement of water and
proactive management approaches, which ultimately support the pro­ contaminants may be tracked, the places at risk of pollution and it can be
tection of water resources. found using geospatial technologies. Sensor technologies can be used to
continuously gather data and measure water quality indicators at spe­
Scope for water quality analysis using geospatial technology, cific places. Data on water quality can be analyzed using machine
sensor technology, and machine learning learning methods to spot trends and patterns. Additionally, machine
learning can be used to forecast changes in water quality, and this can be
One of the main challenges related to collecting contaminated water helpful in developing mitigation strategies and locations that are most
is identifying the threats, analyzing water quality, and producing likely to be impacted by pollution. Processes affecting water quality can
improved water resources management. Water is the most critical need be modeled using geospatial technology, sensor technology, and ma­
of any living organism and essential for supporting the existence of all chine learning methods. We should understand how contaminants affect
animals and individuals. There are definite restrictions on water quality and develop plans to enhance water quality standards.

8
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

description for Fig. 7 is given below the figure. Table 6

Fig. 7 shows the WQI values given for the groundwater tests - G7 and Analysis of Machine Learning Algorithms implemented for assessing water
G21. The spatial guides of the review region are obviously found in the quality (Wang et al., 2017) (Sanya et al., 2022).
places with the most extreme defilement and leachate development from Variables Model Training Validation
the landfill. The WQI standards are related to levels based on the R2 R R.M.S.E. RMSE
threshold level or weightage assigned, and GGIS software is used to R2
analyze the water quality in a specific area. It is GIS, a robust storage and
IWQI Gradient Boosting 0.991 1.481 0.951 2.526
data analysis tool. Using GIS tool, effective water potential zones in a Regression
particular area and targeted sites with hydrological surveys are IWQI Artificial Neural network 0.975 2.482 0.958 2.172
analyzed, and the study was conducted by fixing the longitude and (ANN)
latitude points of the specific location. The government and water WQI Bayesian models 0.85 0.56 0.8 0.65
WQI GA-BPNN 0.932 3.0 0.976 3.13
strategy policies can utilize this information as a solid fundamental WQI RF 0.96 1.37 0.93 3.1
reference in choosing reasonable destinations for groundwater assets. WQI GEP(Gene expression 0.95 14.4 0.89 12.74
The mix of remote detecting with geographic data frameworks (GISs as a programming)
flexible device has expanded the exactness of potential groundwater WQI Ensemble models 0.9 0.4 0.8 0.5
zones. The selection of efficient sensors and processing units is required
to get useful information, and these parameters are sound communica­
tion systems for transferring collected datasets without any loss or de­
viation. Table 5 shows the outcomes with optimal features of various
models of deep learning techniques. The irrigational water quality
indices values are revealed, and 33 % of the samples were not encour­
aged for irrigational purposes mentioned in this study. The most accu­
rate model was predicted using ANN-3F with RMSE for Training and
Testing coefficients as 0.002 and 0.004 and R2 values 0.999 and 0.999.
From Table 6, two metrics, R2 and RMSE values, are used to evaluate
the performance in predicting the water quality index. If the RR.M.S.E.
value is low, it indicates that studies have been directed to look at the
performance of machine learning and A.I. techniques for WQI estima­
tion. These examinations have found that deep learning strategies and
ANN have outperformed traditional machine learning methods, such as
multivariate regression. The model prediction is close to actual WQI
values, and these models can be used to monitor the water quality and
identify the areas where water quality is declining. Implementation of Fig. 8. Region of convergence curve to determine the water quality class.
WQI with a machine learning algorithm is imperative to analyze devi­
ation levels accurately. Big data analysis needs data mining for training investigate areas and shortlisted 249 articles. On this set, we led a bib­
and testing by the different algorithms. The big training data sets need liometric investigation. Then, at that point, we shortlisted 63 pieces for
an efficient algorithm model, and the outcome should be increased at an orderly writing survey. We likewise distinguish the correlative or
the lowest cost. cross-areas utilized in the articles. Table 7 outlines a systematic classi­
From Fig. 8, it is understood that the various studies proposed fication of research queries to steer forthcoming inquiries into respon­
learning strategies that can be utilized to track down the ROC. capa­ sible water quality prediction evolution using Sensor technology, IoT,
bilities with a severe level of precision (Aldhyani et al., 2020). These AI, GIS. These questions fall into three overarching categories: AI
strategies can be utilized to investigate the way of behaving capabilities capability Development and Maturity, Responsible AI implementation
that makes the capabilities diverge. Fig. 9 shows Various performance and Management, and Research Methodology and Validation. Table 7
evaluations of Artificial Intelligence to measure water quality serves as a navigational guide, suggesting potential avenues for future
management. research in Responsible Technologies by furnishing instances of research
The articles included survey articles, peer-reviewed Journals, and questions. (Akbarighatar et al., 2023)
Conference proceedings mentioned in Fig. 10. We confined further to
Conclusions

Water is one of the essential natural resources throughout the world.

Various physio-chemical and biological parameters determine the
excellent quality of water. WWQI is calculated using the standard
equation along with tedious laboratory observation. The machine
learning techniques are applied to reduce the uncertainty of the aggre­
gating process, robustness, and efficiency and lead to water quality
analysis quickly and accurately at a cheaper rate. IoT-based real-time
monitoring data from the various multiple sensors are collected for the
performance analysis of Artificial Intelligence techniques. Water pa­
rameters are identified and validated using big data analysis, and IoT
architecture in water distribution systems gathered from Integrated
sensors is communicated to the cloud for the effective database. The
information collected from various review papers shows that practical
machine learning algorithms are pretty helpful for a better water quality
monitoring system. Machine learning algorithms such as (classification
and regression models) are suggested for the improvement in the control
Fig. 7. WQI maps plot using GIS (Dawood et al., 2022). of water quality. The spatial distributions can estimate the detection and

9
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Fig. 9. Various Artificial Intelligence Methods to Analyze Water Quality Management.

Fig. 10. Various Artificial Intelligence Methods to Analyze Water Quality Management.

prediction of water resources and their quality. Various modeling uses learning, and GIS technologies are encouraging. These technologies can
GRACE, Visual MODFLOW., and ArcGIS 10.2.2, in which the water be utilized to create precise water quality prediction models, according
zones and multiple parameters are identified and predicted. The effec­ to several studies conducted in the past. WQI of the Tigris River in Iraq
tive and practical system for evaluating the nature of groundwater and was predicted using an artificial neural network. This model’s R2 and
its advancement in the utilization of physicochemical boundaries are RMSE values are 0.91 and 0.27, respectively, in the traditional methods
appropriately upheld by GIS methods, multivariate displaying, and of machine learning, and it is. 0.94 and 0.15 in the random forest model.
machine learning. Analytical Hierarchy Process is a central geological to forecast WQI most of the models achieved an R2 value of 0.93 and
technique for identifying ground-level elevation levels, and they per­ RMSE value of 0.17.
formed better boundary conditions and calibration levels. Using these
models, we could ensure support for the sustainable management of CRediT authorship contribution statement
water resources. Overall, numerical outcomes of predicting the water
quality using WQI sensor technology, IoT, machine learning, deep Poornima Jayaraman: . Kothalam Krishnan Nagarajan:

10
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Table 7 Supplementary materials

Example of research questions for future work.
S. Domain Emerging Research Questions Supplementary material associated with this article can be found, in
no Research Areas the online version, at doi:10.1016/j.jjimei.2023.100210.
1 Water Quality Index and i) What strategies can be employed to create more
Prediction inclusive Water Quality Index (WWQI models References
that account for diverse water quality parameters
and the specific attributes of various water Ahmed, U., Mumtaz, R., Anwar, H., Shah, A. A., & Irfan, R. (2019). Efficient water
bodies? quality prediction using supervised. Water, 11, 1–14.
ii) How can we enhance the precision and Akbarighatar, P., Pappas, I., & Vassilakopoulou, P. (2023). International Journal of
dependability of WWQIcomputations, especially Information A sociotechnical perspective for responsible AA.I.maturity models :
when dealing with unconventional water sources Findings from a mixed-method literature review. International Journal of Information
like wastewater and recycled water? iii)How can Management Data Insights, 3, Article 100193. https://doi.org/10.1016/j.
jjimei.2023.100193
we adeptly convey WWQIdata to the general
Aldhyani, T. H. H., Al-Yaari, M., Alkahtani, H., & Maashi, M. (2020). Water quality
public and stakeholders, fostering awareness and
prediction using artificial intelligence algorithms. Applied Bionics and Biomechanics,
encouraging practices that promote water
2020. https://doi.org/10.1155/2020/6659314
conservation?iv) How can we incorporate real- Almetwally, S. A. H., Hassan, M. K., & Mourad, M. H. (2020). Real time Internet of
time data from sensor networks and IoT Things (IoT) based water quality management system. Procedia CC.I.R.P, 91,
infrastructure into predictive models to improve 478–485. https://doi.org/10.1016/j.procir.2020.03.107
the timeliness and accuracy of water quality Arabameri, A., Pal, S. C., Rezaie, F., Nalivan, O. A., Chowdhuri, I., Saha, A., Lee, S., &
forecasts? Moayedi, H. (2021). Modeling groundwater potential using novel GIS-based
2 Sensor Technology i)What approaches can be taken to create sensors machine-learning ensemble techniques. Journal of Hydrology: Regional Studies, 36,
with heightened sensitivity, selectivity, and Article 100848. https://doi.org/10.1016/j.ejrh.2021.100848
durability capable of detecting an extended array Azpilicueta, L., Aghzout, O., Alejos, A.V., Falcone, F., Analysis of low power wide area
of water quality parameters, encompassing network wireless technologies in smart agriculture for large-scale farm monitoring
emerging contaminants and trace chemicals? ii) and tractor communications, 187 (2022).
In what ways can we advance the energy Bates, H., Pierce, M., Benter, A., LoRaWAN-based IoT sensor network, (2021) 1–14.
Bell, C. (2013). Planning Wireless Sensor Networks, Beginning Sensor Networks with Arduino
efficiency and lifespan of sensors to mitigate
Raspberry Pi (pp. 309–333). Springer Nature. https://doi.org/10.1007/978-1-4302-
environmental consequences and lower
5825-4_9
maintenance expenditures? iii)How can we Bui, D. T., Khosravi, K., Tiefenbacher, J., Nguyen, H., & Kazakis, N. (2020). Improving
design sensor networks that are scalable, prediction of water quality indices using novel hybrid machine-learning algorithms.
economically viable, and adaptable to a variety of Science of the Total Environment, 721, Article 137612. https://doi.org/10.1016/j.
water environments, including remote and scitotenv.2020.137612
challenging locations? Chen, K., Chen, H., Zhou, C., Huang, Y., Qi, X., Shen, R., Liu, F., Zuo, M., Zou, X.,
3 Machine Learning i)What strategies can be employed to devise Wang, J., Zhang, Y., Chen, D., Chen, X., Deng, Y., & Ren, H. (2020). Comparative
groundbreaking machine learning algorithms to analysis of surface water quality prediction performance and identification of key
manage extensive and intricate water quality water parameters using different machine learning models based on big data. Water
datasets, accommodating challenges like missing Research, 171, Article 115454. https://doi.org/10.1016/j.watres.2019.115454
data, outliers, and nonlinear relationships? ii) Dawood, A.S., Jabbar, M.T., Al-tameemi, H.H., Baer, E.M., Application of water quality
How can machine learning be seamlessly index and multivariate statistical techniques to assess and predict of groundwater
quality with aid of geographic information system, 23 (2022) 189–204.
integrated with other data analysis techniques,
Demetillo, A. T., Japitana, M. V., & Taboada, E. B. (2019). A system for monitoring water
like statistical modeling and data mining, to
quality in a large aquatic area using wireless sensor network technology. Sustainable
enrich comprehension of water quality patterns
Environment Research, 1, 10–12. https://doi.org/10.1186/s42834-019-0009-4
and trends? iii)How can we fashion machine Dilmi, S., & Ladjal, M. (2021). A novel approach for water quality classification based on
learning models that offer Transparency in the the integration of deep learning and feature extraction techniques. Chemometrics and
decision-making process and facilitate the Intelligent Laborary Systems, 214, Article 104329. https://doi.org/10.1016/j.
interpretability of results? chemolab.2021.104329
4 GIS(Geographical i)What methodologies can be employed to Dohare, D., Deshpande, S., & Kotiya, A. (2014). Analysis of ground water quality
information system) advance GGIS tools and techniques, facilitating parameters: A review www.isca.me. Research Journal of Engineering Sciences, 3,
the seamless integration, analysis, and 2278–9472.
visualization of spatial data pertinent to water Duan, Y., Edwards, J. S., & Dwivedi, Y. K. (2019). International Journal of Information
quality, encompassing aspects like land use, Management Artificial intelligence for decision making in the era of Big Data –
topography, and pollution sources? ii)How can Evolution , challenges and research agenda. International Journal of Information
Management, 48, 63–71. https://doi.org/10.1016/j.ijinfomgt.2019.01.021
GGIS be effectively harnessed to pinpoint
Watershed, E. E. L., Wang, X., Zhang, F., & Ding, J. (2017). Evaluation of water quality
hotspots of water pollution, potential sources of
based on a machine learning algorithm and water quality index for the. Scientific
contamination, and regions susceptible to water
Reports, 1–18. https://doi.org/10.1038/s41598-017-12853-y
quality degradation? iii)What approaches can be El Bilali, A., & Taleb, A. (2020). Prediction of irrigation water quality parameters using
taken to formulate GIS-based decision-support machine learning models in a semi-arid environment. Journal of the Saudi Society of
systems tailored for water resource management, Agricultural Sciences, 19, 439–451. https://doi.org/10.1016/j.jssas.2020.08.001
pollution control, and regulatory compliance? Ewaid, S. H., Abed, S. A., & Kadhum, S. A. (2018). Predicting the Tigris River water
quality within Baghdad, Iraq by using water quality index and regression analysis.
Environmental Technology & Innovation, 11, 390–398. https://doi.org/10.1016/j.
Conceptualization, Data curation, Formal analysis, Supervision, Writing eti.2018.06.013
Ewaid, S. H., Kadhum, S. A., Abed, S. A., & Salih, R. M. (2019). Development and
– review & editing. Pachaivannan Partheeban: Conceptualization, evaluation of irrigation water quality guide using II.W.Q.G.V.1 software: A case
Data curation, Formal analysis, Methodology, Resources, Supervision, study of Al-Gharraf Canal, Southern Iraq. Environmental Technology & Innovation, 13,
Validation, Writing – original draft, Writing – review & editing. Valli­ 224–232. https://doi.org/10.1016/j.eti.2018.12.001
GG.S. (2013). Water quality index with missing parameters. International Journal of
devi Krishnamurthy: Data curation, Supervision, Writing – review & Research in Engineering and Technology, 02, 609–614. https://doi.org/10.15623/
editing. ijret.2013.0204035
Ghose, M. K., Dikshit, A. K., & Sharma, S. K. (2006). A GIS based transportation model for
solid waste disposal - A case study on Asansol municipality. Waste Management, 26,
Declaration of Competing Interest
1287–1293. https://doi.org/10.1016/j.wasman.2005.09.022
Herath, HH. M. K. K. M. B., & Mittal, M. (2022). Adoption of artificial intelligence in
The authors declare that they have no known competing financial smart cities : A comprehensive review. International Journal of Information
interests or personal relationships that could have appeared to influence Management Data Insights, 2, Article 100076. https://doi.org/10.1016/j.
jjimei.2022.100076
the work reported in this paper. Hussein, E. A., Thron, C., Ghaziasgar, M., Bagula, A., & Vaccari, M. (2020). Groundwater
prediction using machine-learning tools. Algorithms, 13, 1–16. https://doi.org/
10.3390/a13110300

11
P. Jayaraman et al. International Journal of Information Management Data Insights 4 (2024) 100210

Jha, D. K., Devi, M. P., Vidyalakshmi, R., Brindha, B., Vinithkumar, N. V., & Journal of Information Management Data Insights, 1, Article 100050. https://doi.org/
Kirubagaran, R. (2015). Water quality assessment using water quality index and 10.1016/j.jjimei.2021.100050
geographical information system methods in the coastal waters of Andaman Sea, Sn, J., Muramthookil Thomas, G., Rohith Raj, R. V., Masetti, A., Tammana, A.,
India. Marine Pollution Bulletin, 100, 555–561. https://doi.org/10.1016/j. Motheram, M., & Gutlapalli, N. C. (2020). Assessment of water quality Index and
marpolbul.2015.08.032 study of the impact of pollution on the rivers of Kerala. Materials Today: Proceedings,
Kalini, Z., International Journal of Information Management A SEM-neural network 43, 3447–3451. https://doi.org/10.1016/j.matpr.2020.09.084
approach for predicting antecedents of m-commerce acceptance, 37 (2017) 14–24. Sun, W., Xia, C., Xu, M., Guo, J., & Sun, G. (2016). Application of modified water quality
https://doi.org/10.1016/j.ijinfomgt.2016.10.008. indices as indicators to assess the spatial and temporal trends of water quality in the
Kar, A. K., & Varsha, P. S. (2023). A review and research agenda. International Journal of Dongjiang River. Ecological Indicators, 66, 306–312. https://doi.org/10.1016/j.
Information Management Data Insights. , Article 100176. https://doi.org/10.1016/j. ecolind.2016.01.054
jjimei.2023.100176 Sutadian, A. D., Muttil, N., Yilmaz, A. G., & Perera, B. J. C. (2016). Development of river
Kothari, N., Shreemali, J., Chakrabarti, P., & Poddar, S. (2021). Design and water quality indices—A review. Environmental Monitoring and Assessment, 188,
implementation of IoT sensor based drinking water quality measurement system. 1–29. https://doi.org/10.1007/s10661-015-5050-0
Materials Today: Proceedings. https://doi.org/10.1016/j.matpr.2020.12.1142 Swamee, P. K., & Tyagi, A. (2000). Describing water quality with aggregate index.
Lamare, R.E., Singh, O.P., Application of Ccme water quality index in evaluating the Journal of Environmental Engineering, 126, 451–455. https://doi.org/10.1061/(ASCE)
water quality status in limestone mining area of, 10 (2016) 149–154. 0733-9372(2000)126:5(451)
Liu, P., Wang, J., Sangaiah, A., Xie, Y., & Yin, X. (2019). Analysis and prediction of water Thinsungnoen, M., An integrated system of geographic information and water quality :
quality using LL.S.T.M.Deep neural networks in IoT environment. Sustainability, 11, Lamtakong river, 10 (2019) 152–159. https://doi.org/10.12720/jait.10.4.152-159.
2058. https://doi.org/10.3390/su11072058 Tiwari, A. K., Singh, A. K., & Mahato, M. K. (2018). Assessment of groundwater quality of
Lumb, A., Sharma, T. C., & Bibeault, J.-F. (2011). A review of genesis and evolution of Pratapgarh district in India for suitability of drinking purpose using water quality
water quality index (WWQI and some future directions. Water Quality, Exposure and index (WWQI and GGIStechnique. Sustainable Water Resources Management, 4,
Health: Purpose and Goals, 3, 11–24. https://doi.org/10.1007/s12403-011-0040-0 601–616. https://doi.org/10.1007/s40899-017-0144-1
Mancuso, P., Piccialli, V., & Sudoso, A. M. (2021). A machine learning approach for Tuna, G., Arkoc, O., & Gulez, K. (2013). Continuous monitoring of water quality using
forecasting hierarchical time series. Expert Systems with Applications, 182, Article portable and low-cost approaches. International Journal of Distributed Sensor
115102. https://doi.org/10.1016/j.eswa.2021.115102 Networks, 2013. https://doi.org/10.1155/2013/249598
Najah Ahmed, A., Binti Othman, F., Abdulmohsin Afan, H., Khaleel Ibrahim, R., Ming Uddin, M. G., Nash, S., & Olbert, A. I. (2021). A review of water quality index models and
Fai, C., Shabbir Hossain, M., Ehteram, M., & Elshafie, A. (2019). Machine learning their use for assessing surface water quality. Ecological Indicators, 122, Article
methods for better water quality prediction. Journal of Hydrology, 578. https://doi. 107218. https://doi.org/10.1016/j.ecolind.2020.107218
org/10.1016/j.jhydrol.2019.124084 Venkatachalam, P., & Ray, S. (2022). How do context-aware artificial intelligence
Nampak, H., Pradhan, B., & Manap, M. A. (2014). Application of GGISbased data driven algorithms used in fitness recommender systems ? A literature review and research
evidential belief function model to predict groundwater potential zonation. Journal agenda. International Journal of Information Management Data Insights, 2, Article
of Hydrology, 513, 283–300. https://doi.org/10.1016/j.jhydrol.2014.02.053 100139. https://doi.org/10.1016/j.jjimei.2022.100139
Nayak, J. G., Patil, L. G., & Patki, V. K. (2020). Development of water quality index for Wang, X., Zhang, F., & Ding, J. (2017). Evaluation of water quality based on a machine
Godavari River (India) based on fuzzy inference system. Groundwater for Sustainable learning algorithm and water quality index for the Ebinur Lake Watershed, China.
Development, 10, Article 100350. https://doi.org/10.1016/j.gsd.2020.100350 Scientific Reports, 7, 1–18. https://doi.org/10.1038/s41598-017-12853-y
Nayar, R., Dr. (2020). Assessment of water quality index and monitoring of pollutants by Wu, Z., Zhang, D., Cai, Y., Wang, X., Zhang, L., & Chen, Y. (2017). Water quality
physico-chemical analysis in water bodies: A review. International Journal of assessment based on the water quality index method in Lake Poyang: The largest
Engineering Research, 9, 178–185. https://doi.org/10.17577/ijertv9is010046 freshwater lake in China. Scientific Reports, 7, 1–10. https://doi.org/10.1038/
Pasika, S., & Gandla, S. T. (2020). Smart water quality monitoring system with cost- s41598-017-18285-y
effective using IoT. Heliyon, 6, e04096. https://doi.org/10.1016/j.heliyon.2020. Wu, Z., Lai, X., & Li, K. (2021). Water quality assessment of rivers in Lake Chaohu Basin
e04096 (China) using water quality index. Ecological Indicators, 121, Article 107021. https://
Roseela, J. A., Godhavari, T., Narayanan, R. M., & Madhuri, P. L. (2020). Design and doi.org/10.1016/j.ecolind.2020.107021
deployment of IoT based underwater wireless communication system using Xu, X., Liu, Y., Liu, S., Li, J., Guo, G., & Smith, K. (2019). Real-time detection of potable-
electronic sensors and materials. Materials Today: Proceedings, 45, 6229–6233. reclaimed water pipe cross-connection events by conventional water quality sensors
https://doi.org/10.1016/j.matpr.2020.10.586 using machine learning methods. Journal of Environmental Management, 238,
Sanya, WW.M., Alawi, MM.A., Eugenio, I., Design and development of smart water 201–209. https://doi.org/10.1016/j.jenvman.2019.02.110
quality monitoring system using IoT design and development of smart water quality Yang, R., Yu, L., Zhao, Y., Yu, H., Xu, G., Wu, Y., & Liu, Z. (2019). International Journal
monitoring system using IoT, (2022). https://doi.org/10.31695/IJASRE.2022.8.3.1. of Information Management Big data analytics for fi nancial Market volatility
Shakhari, S., & Banerjee, I. (2019). A multi-class classification system for continuous forecast based on support vector machine ⋆. International Journal of Information
water quality monitoring. Heliyon, 5, e01822. https://doi.org/10.1016/j. Management. https://doi.org/10.1016/j.ijinfomgt.2019.05.027, 0–1.
heliyon.2019.e01822 Yeh, H. F., Cheng, Y. S., Lin, H. I., & Lee, C. H. (2016). Mapping groundwater recharge
Sharma, R., Kumar, A., & Chuah, C. (2021). International Journal of Information potential zone using a GGISapproach in Hualian River, Taiwan. Sustainable
Management Data Insights Turning the blackbox into a glassbox : An explainable Environment Research, 26, 33–43. https://doi.org/10.1016/j.serj.2015.09.005
machine learning approach for understanding hospitality customer. International

12