EFFICIENT WATER QUALITY ANALYSIS AND PREDICTION USING
MACHINE LEARNING
TEAM ID: PNT2022TMID39608
TEAM MEMBERS:
1. S.SUDHARSAN
2. N.SIDDHARTH
3. F.SUHAIL
4. M.VIGNESH
PROBLEM STATEMENT
Water is considered as a vital resource that affects the various aspects of human health and
lives. The quality of water is a major concern for people living in the urban areas.
Quality of water serves as a powerful environmental determinant and a foundation for the
prevention and control of water borne diseases.
However predicting the urban water quality is a challenging task since the water quality varies
in urban spaces non linearly and depends on multiple factors, such as water usage patterns,
temperature, pH, total amount of Dissolved solids, etc.
So it has created a necessity to determine the quality of water samples so as to determine and
detect the contaminants present in advance.
PURPOSE
The main purpose of this project is :
◦ To create a safe and healthy environment by preventing the water borne
diseases among normal public by testing the Quality of the sources of water
they consume.
◦ Develop a Machine Learning model that predicts the Water Quality Index of
any kind of water samples.
◦ Create a model such that it should trust and belief among the customers who
are using the system to predict the water sample they utilize for their needs.
PROJECT OBJECTIVES
To predict the Water Quality Index (WQI) for any kind of water samples using
the Machine Learning Regression Algorithms based on the important
parameters like temperature, pH, total amount of Dissolved and Suspended
solids, etc.
To provide any information regarding the purification technique recommended
based on the impurities present in the water sample tested.
To predict the Water Quality Classification (WQC) using Machine Learning
Classification algorithms.
SCOPE OF THE PROJECT
Create a model in such a way that it can be used by all kinds of people like:
Water Quality Testing Agencies
Private and Public Laboratories
Various industries like Textile industry, Cotton industry, etc.
General Public for testing the water they consume for household purposes.
It should also possess the ability to sustain the changes by testing any kind of water samples
and also to have broader scope of integrating the model with Future Technologies also.
LITERATURE REVIEW
PREDICTING OF WATER QUALITY USING MULTIVARIATE LINEAR REGRESSION :
◦ This model is used to measure the Biological Oxygen Demand (BOD), Chemical Oxygen
Demand (COD) based on the following four parameters: Temperature, pH, total suspended solids
and dissolved solids.
◦ This approach proposed that the deterministic and multivariate linear regression models were
used to speed up the process of predicting the water quality.
◦ As the dataset is considered as time based series, so it is likely to have non linear relationship.
So, the performance of this approach is expected to be poor, with large prediction error.
APPLICATION ADAPTIVE NEURO FIZZY INFERENCE SYSTEM (ANFIS):
◦ Many studies have proven that ANFIS, which can integrate linear and non linear relationships
hidden in the dataset, is a better option in predicting the effluent water quality and also shows
that the ANFIS model works better than the ANN (Artificial Neural Network) model in
predicting the Dissolved Oxygen content in the water sample to be tested.
◦ The model predicts the water quality based on Eight Parameters to predict the total
phosphorous and total nitrogen content in the water sample. The model could accurately
formulate the hidden relationships and correlation analysis can improve the prediction
accuracy.
◦ The disadvantage is that this proposed model approach requires that the size of training dataset
should not be less than the number of training parameters and if the correlation between the
data in the dataset are weak then it generates out of range errors.
SOLUTION ARCHITECTURE
TECHNOLOGY ARCHITECTURE
PROCESS INVOLVED
METHODOLOGIES APPLIED IN EACH PHASES
DFD LEVEL 0 DIAGRAM
DFD LEVEL 1 DIAGRAM
PROJECT FLOW
Dataset Model Building
Collect the Dataset or create the Dataset. Training and Testing the Model.
Data Preprocessing.
Evaluation of Model.
Importing the libraries.
Application Building
Importing the Dataset.
Create an HTML file.
Checking for Null Values.
Build a Python code.
Data Visualization.
Taking care of Missing Data.
Label Encoding.
Splitting the Data into Train and Test.
DATASET
◦ It contains the different samples from different Indian states and has 7 significant parameters namely
1. Dissolved Oxygen
2. pH
3. Conductivity
4. Biological Oxygen Demand
5. Nitrate
6. Fecal Coliform
7. Total Coliform
DATA PREPROCESSING
◦ It is an important phase in the Data analysis to improve the data quality.
◦ The WQI has been calculated from the most significant parameters of the dataset.
◦ Then the water samples have been classified on the basis of the WQI values.
MODEL BUILDING
◦ Building an ML model requires splitting of Data into two sets such as
1. Training Set
2. Testing Set
A set of supervised (for labelled) and unsupervised (for unlabeled data) algorithms are available to be chosen depending on the
Nature of Data and business outcome to predict.
We have chosen “Random Forest Regression” Algorithm as the model as we are predicting a numeric data WQI value based on
the various numeric basic Water Quality Parameters.
APPLICATION BUILDING
◦ We are using various Application Frameworks for creating an Interactive and User Friendly application so that the users can
predict the Water Quality Index (WQI) value based on the inputs given by the user on the interface.
◦ Application Framework:
◦ It is a software library that provides a fundamental structure to
support the development of applications for a specific environment.
◦ Application Frameworks Used:
1. Flask (Light Weight WSGI web application framework.
2. Anaconda Navigator
*(Distribution of R and Python Programming languages for scientific computing.
Output
◦ User Interface Page
WHY TO FIND WATER QUALITY PAGE
Water Quality Determining Parameters
◦ pH
◦ Dissolved Oxygen
◦ Biological Oxygen Demand:
◦ Coliform:
◦ Conductivity:
◦ Nitrate:
WQI INDEX VALUE ANALYSIS TO
CREATE AN AWARENESS
GOOD WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ INPUT :
GOOD WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ PREDICTION:
FAIR WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ INPUT:
FAIR WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ PREDICTION:
POOR WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ INPUT:
POOR WATER QUALITY SAMPLE DATA
AND PREDICTION
◦ PREDICTION:
THANK YOU
