IJESC, August 2021 28625 http:// ijesc.org/ ISSN 2321 3361 © 2021 IJESC Identifying insights on diabetes using Knowledge Graph Somesh Kumar Sahu 1 , Pradeepta Mishra 2 , Dr. Shinu Abhi 3 REVA Academy for Corporate Excellence, REVA University Bangalore, India Abstract: Diabetes affects nearly 11.5% of the Indian population, and globally almost 18% of the total population.High glucose levels cause the early signs of diabetes, if not treated, will lead to severe consequences. Managing diabetes with self-administered precautions is crucial. There are vast amounts of information on managing diabetes available across the web through blogs and websites. However, converting these data into meaningful knowledge isn’t always straightforward. This paper describes how diabetes-related informationavailable in the web can be transformed into ‘knowledge graphs’. The knowledge graph identifies facts, reduces noise, and determines what information is missing from an extraction graph. In building the knowledge graph, the data sources play a significant role. The integrity of the sources is always a challenge. This paper proposes combining web crawling techniques withNLTK to process unstructured data, add transformations, and use knowledge graphs to construct relational graphical representations. This study will help the researchers, hospitals, diabetes experts, and patients to learn how to maintain their diabetes levels and control them. Keywords:Text Mining, Natural Language Processing, NLTK, Knowledge Graph, Network X. I. INTRODUCTION Diabetes ranks as the top five leading causes of death globally, with almost 18% of the population affected[1]. 11.5% of Indian adults aged 45 to 59 self-report to have diabetes; older adults (14%) have a higher prevalence rate than younger adults (9%).Insulin, a hormone made by the pancreas, helps glucose from food get into the cells to be used for energy.The body becomes diabetic when it cannot produce enough insulin to convert sugar into energy. Diabetic patients need to follow a strict regimen; regular tests, dietary control etc. become part of their everyday routine. If ignored, cardiovascular problems are more likely to develop when diabetes is presentAccording to clinical studies conducted in urban areas, where type 2 diabetes is very common, proof of early detection reduces its frequency.To lead a healthy lifestyle, the tests need to be taken at regular intervals[2]. By using clinical tests, one will be able to manage diabetes. Today there are many applications and devices helpingpeople maintain a balanced and healthy lifestyle. Other than the devices, most of us get information from websites too.To be fair, it is an important point to know what all the current highlights are, as well as what type of concepts we have, since it is becoming challenging to consume valid and accurate data. This paper focuses on how one can get new subtleties by combining machine learning and knowledge graphs. The NLP system examines the vast amount of data available in the form of text into useful information. This combines phonetics, figurative structures, and man-made ideas of sub-fields. Python provides several libraries for processing ordinary language, including NLTK, Text Blob, Beautiful Soup, and spaCy. Graphics illustrate relationships between parts and cover an incredible amount of space, absorbing a great deal of information. The main challenge of a knowledge-based system lies in accurately representing the knowledge. Graphs of knowledge have attracted both academic and industry attention. For web applications, they are widely used as a knowledge representation technique. Retrieving patterns and representing knowledge is critical to knowledge discovery. The graph model of presenting context information in the text is different from other pieces of knowledge’s representation, such as production rules, and has been proposed to address complex relevance issues in a more user-friendly way [3]. Diabetes patients can get healthy when they know how to take appropriate preventive measures to care for themselves daily. Finding the correct information or searching for relevant information is time-consuming because there are so many resources available. The purpose of this study is to support those who believe that Internet research can provide real-time access to information. They need to spend several hours getting the information they need, but with techniques like knowledge graphs, this can be accomplished in a short time. II. UNDERSTANDING DIABETES Today, more than a billion people live with diabetes worldwide. Diabetes is a long-term condition characterized by the inability of the body to produce insulin or the inability to use the insulin that it makes [4]. Since 2000, the number of diabetes patients have tripled to 463 million, up from 151 million in 2000 (near to the WHO estimate of 150 million at the time). In the coming years, diabetes will continue to impact the global economy at increasing levels.Type 1, type 2, and gestational diabetes mellitus are the most common types of diabetes. Diabetes mellitus (DM) is diagnosed and treated by measuring glycated hemoglobin (HbA1c) in whole blood (B-HbA1c), or venous plasma glucose concentrations (P-glucose)[5]. Type 1 is the most frequent reason for diabetes in youth, but it can also develop at any stage of life. It is possible for people living with type 1 diabetes to live solid, satisfying lives if they have a continuous supply of insulin and physical training. Type 2Most people who have diabetes (90%) are suffering from type 2. Guidelines, support, and following healthy lifestyles are effective ways of managing it. There is growing evidence that type 2 diabetes can be treated. Research Article Volume 11 Issue No.08