Result Analysis Chatbot

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

Result Analysis Chatbot

Prof. Bharath Bharadwaj B S1 , Mr. Ayush J2 , Mr. Rajesh S3, Mr.Mohammed Sameer4

1Assistant Professor, Dept. of Computer Science and Engineering, Maharaja Institute of Technology, Thandavapura

234Students, Dept of Computer Science and Engineering, Maharaja Institute of Technology, Thandavapura ***

Abstract - The undertaken project is an offline smart chatbot that makes educational result analysis easy. Running in offline mode, it enables the user to engage in natural language to query student performance data. It processes structured documents such as PDFs or CSVs toprovide insights in the form of subject-wise marks, pass/fail statistics, toppers, and overall aggregates. With Python, it makes use of libraries such as pandas for handling data and pdfplumber for content extraction. Utilizing simple NLP methods, the chatbot interprets questions and gives correct feedback. Its structured nature allows smooth updates and structure integration, hastening result study, making it secure, as well as optimizing it for teaching staff, pupils, and admins

Key Words: ResultAnalysisChatbot,offlinechatbot,natural language processing (NLP), PDF parsing, CSV parsing, PyPDF2,pdfplumber,pandas,Python

1.INTRODUCTION

Result Analysis Chatbot is a revolutionary offline web application designed to streamline and mechanize the processofacademicresultanalysis.Itwasconstructedusing Python and the Flask web framework, the system allows schools to process efficiently student outcome data and create informative performance reports no coding required aninternetconnection. Itsupportsvariousinput formatslikePDF,CSV,andExcelfiles,andpullsoutrelevant learning information such as student names, subject-wise marks,grades,pass/failstanding,CGPA,andperformance. Afterexaminingthedata,thechatbotgeneratesacomplete Excelreportcontainingthoroughstatistics,alistoftoppers, and graphical figures such as graphs and charts to better understandcomparison.Thisprojectaddressesthetypical issues of manual result analysis by offering a time-saving, offline,anduser-friendlysolutionthatsavestime,reduces errors,andenhancesdataavailabilityforteachers,students, andadministratorsaswell.

1.1 Objective

ThemainobjectiveoftheResultAnalysisChatbotprojectis to develop an offline, intelligent web application that automates and simplifies academic result analysis. The systemcanacceptresultdatainformatssuchasPDF,CSV, andExcel,andefficientlyextractkeyacademicdetailslike studentnames,marks,subjects,andgrades.Byincorporating Natural Language Processing (NLP), the chatbot enables

userstointeractwiththesystemusingnatural,human-like queries,makingithighlyuser-friendlyandaccessibleeven fornon-technicalusers.

1.2 Motivation to take up the project

ThemotivationtobuildtheResultAnalysisChatbot arises from the increasingdemand for smarter, faster,andmore efficienttoolswithintheacademiccommunity,particularly for assessing student performance. In many schools and colleges, administrators and teachers invest considerable timeinreviewingresults,calculatingperformancemetrics, and identifying trends, which can slow down decisionmaking and increase the risk of human error especially whendealingwithlargevolumesofdata.Whilesomedigital toolsdoexist,theyareoftentoocomplexfornon-technical usersorheavilyreliantoninternetconnectivity,whichmay notalwaysbeavailableordependableineveryeducational environment.

1.3

Challenges to be addressed

The Result Analysis Chatbot addresses several major challenges commonly encountered in academic result processing.Manyinstitutionscontinuetorelyonmanualor semi-automated systems, making the process slow and error-prone. Existing tools often require a stable internet connection,whichmaynotbeavailableinruralorremote areas.Moreover,thesesystemsaregenerallycomplexand not suitable for non-technical users, lacking intuitive and user-friendlyinterfaces.

TheytypicallydonotsupportmultiplefileformatslikePDF, CSV, and Excel, limiting flexibility in data handling. Most availablesolutionsalsofailtoprovidemeaningfulgraphical insightssuchaschartsandgraphs,whichareessentialfor clearanalysis.Furthermore,thelackofinteractiveNatural Language Processing (NLP) features reduces accessibility whenretrievingspecificdata.Scalabilityisanotherconcern, as many systems struggle to manage large datasets efficiently.Thisprojectovercomestheseissuesbyofferinga secure, offline, NLP-powered, and easy-to-use platform tailoredforeducationalenvironments

1.4

Technology stack

Python is a high-level, general-purpose programming languagethatiseasytoreadandwrite,simple,andversatile.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

It is a multi-paradigm language that supports objectoriented, procedural, and functional programming, and is predominantly used in web development, data analysis, artificial intelligence, machine learning, and automation. Withitslargestandardlibraryandavastcommunity,Python provides an ideal platform for rapid application development.Itssimplesyntaxmakesiteasyforbeginners to learn and use, while being adequate for large software projects.IntheResultAnalysisChatbot,Pythonservesasthe foundationforbackendlogic,fileparsing,dataprocessing, and integration of Natural Language Processing (NLP) procedures.

FlaskisasmallandnimblePythonwebframework.Itstrives to make starting web development a quick and painless process,butstillleavethepotentialtoscaleuptocomplex applications.FlaskisWSGI-compatibleandusestheJinja2 templating engine to template dynamic web pages. It has built-in support for routing, handling HTTP requests, and session handling, but otherwise is opinion-free leaving developers free to use their own tools and libraries for database access, form handling, and other tasks. In this project, Flask is used to build the web interface to the chatbot,whereuserscanuploadresultfiles,sendmessages to the chatbot over HTTP routes, and view analyzed results allwithinasmooth,responsive,andoffline-friendly interface.

1.5 Literature survey

Significant advancements have been made in the field of braintumordetectionandclassification,withseveralstudies contributingtothedevelopmentofimprovedmethodologies.

OnenotablestudybyMr.P.Nareshetal.CreatingaFlaskbasedChatbottoProcessUniversityQueriesusingSpacyand TensorFlow Mr. P.Nareshetal.developeda chatbotbuilt withFlask,Spacy,andTensorFlowtohandlecollege-related queries.Thechatbotusesanintentclassificationmodelwith preconfigured patterns and responses in JSON format. Utilizingdeeplearningandnaturallanguageprocessing,the systemconvertsuserqueriesintotherespectiveintentsand responds accordingly. The web interface allows for interactiveinteraction,whilecompatibilitywitheducational data provides rapid and meaningful information. Though easy to design, the chatbot offers stable performance and scalability, serving as a good assistant in institutional applications.

UdayBhamreetal. designedaprivate,personalchatbotthat canbeusedtointeractwith documentfileslikePDF,TXT, andCSVwithoutuploadingtotheinternet.Thearchitecture employs open-source Large Language Models (LLMs) and buildsalocalknowledgebaseusingtechniqueslikesemantic indexingandChroma vector stores.Thesolutionprovides 100%dataprivacy,solvingproblemsofcommercialchatbot data leaks like those of ChatGPT. The solution provides private querying of confidential documents on local

machineswithsecurityand performanceinenvironments requiringstrictdataconfidentiality.

SiddharthVermaetal.CHAI:AChatbotAIforTask-Oriented Dialogue with Offline Reinforcement Learning Siddharth Verma et al. designed CHAI, an offline reinforcement learningchatbotforgoal-directeddialogue.Whilelearning fromrealconversationscanbechallenging,CHAIlearnsto maximizegoal-directedconversations

StergiosSozosandDimitriosZarrisproposedanAIchatbot for learning with dual function: a Teacher Assistant for curriculum-alignedexamgenerationandaStudyBuddyfor subject-specific guidance rendering. The system explores promptengineeringtechniqueslikepersonatemplates,fewshot,andchain-of-thoughtpromptingtofacilitateimproved responseaccuracy.Withlargelanguagemodelintegration, the chatbot provides assistance in subjects like biology, math,andphysics.Theexperimentmeasuresperformance onsuchmetricsasrelevanceandcompleteness,showingthe effectofpromptdesignonoutcomequality.Thesolutionhas great potential in AI-assisted learning guidance and personalizedlearning.

Shivom Agarwal et al introduced a domain-generalizable chatbot system based on reinforcement learning and NLP that can be transferred across domains such as customer service, finance, and healthcare. The most important innovations are emotion detection, multilingual voice interaction, offline capability, and customer preference modelingbasedondynamicsatisfactionscores.

The system improves user experience by dynamically adjusting conversations and mimicking human-like interactions.Thesystemalsofocusesonethicalaspectssuch as fairness and data privacy. The flexibility and domain transferabilityofthemodelreflectitsuseasacontemporary conversational AIsystemforservice qualityenhancement andpersonalization.

2. EXISTING SYSTEM

Modernstudentresultanalysissystemsaretypicallybuilton simple CRUD operations (Create, Read, Update, Delete) on PHP or Java-based programming languages with MySQL database support. Most software solutions are web-based withrole-basedaccess(admin,student)supportwithlogin authentication.Mostsuchsystemsdonothavewidespread real-time integration with learning environments and lack supportforliveupdatesfromtestsorexams.

Further,modernsolutionsbarelyscaleupwellandinvolvea tremendousamountofmanualeffortinsupportingadditional departments, courses, or assessment frameworks. User interfaces are typically heterogeneous, with some systems having only command-line interfaces, thus becoming cumbersometonon-technicalusers.

Research

2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

3. PROPOSED SYSTEM

The Result Analysis Chatbot is intended to help schools manage academic information in a more efficient manner. Manualanalysisgetstediousandcumbersomewithagreater numberofstudentsandperformanceresults.Theapplication makes it much easier, and as it's automated, anyone can manageit.

Oneofthestrongestfeaturesisthatitworksoffline,unlike mostapps,whichrequireinternetconnectivity.This offers enhanced privacy, security, and convenience, especially wherethereisweakornointernetconnectivity.Uniqueto this project is the implementation of Natural Language Processing(NLP)onanofflinewebapplication.Anybodycan quiteeasilyusethesystembymerelyaskingquestionsusing naturallanguage,forinstance,"Showtopthreestudentsin Science"or"Howmanystudentsfailed?"Itisquitesimplefor anyonetousewithoutanytechnicalskill.

4. SYSTEM ARCHITECTURE

The offline Result Analysis Chatbot system architecture is builtwiththeFlaskframework.Theusercommunicateswith thesystemviaalocalwebinterfaceofferedbytheFlaskWeb UIwithintheirbrowser.Ifa useruploadsafileorsendsa query, the Flask server's Controller processes the request and forwards it to the AppState Manager. The AppState Manager is the centerpiece that orchestrates various modules within it to extract the data. The Text Extractor moduleextractsacademicdatafromuploadedPDForDOCX files,andrawdataextractedbytheTextExtractormoduleis processedagainbytheSubject/StudentExtractortoextract studentnames,subjects,andscores.

The Excel Report Generator produces a formatted Excel reportafterstructuringthedata,anditissavedasanoutput file.Also,theChatbotQueryHandlerenablesuserstopose questionsinnaturallanguage,like"Showtopthreestudents inScience"or"Howmanystudentsfailed?",anditretrieves correctanswersfromtheprocesseddata.Thewholesystem operates offline, providing privacy, security, and accessibility, particularly in regions with poor internet connectivity.

5. IMPLEMENTATION AND RESULT

The deployment of the Result Analysis Chatbot involves buildinga modular offline web applicationusing theFlask framework in Python. The user interface is crafted with Flask’sJinja2templatingengine,HTML,CSS,andJavaScript, allowinguserstouploadresultdocuments,viewgenerated Excelreports,andinteractwiththechatbotthroughnatural language queries. The backend features a controller that handlesincomingrequestsandroutesthemappropriately. TheAppStateManagercoordinatestheoveralldataflowand session state of the application. A Text Extractor module, utilizinglibrariessuchasPyMuPDForpython-docx,extracts rawtextfromuploadedPDForDOCXfiles.Thisextractedtext is processed by the Subject/Student Extractor, which uses string parsing and regular expressions to identify student names,subjects,andmarks.

The structured data is then used by the Excel Report Generator, built with libraries like pandas or openpyxl, to create well-formatted Excel reports that include rankings, performancesummaries,andpass/failstatistics.TheChatbot QueryHandler,employingbasicNaturalLanguageProcessing techniquesthroughlibrarieslikespaCyorrule-basedparsing, interpretsuser queriesandretrievesaccurate information fromthedataset.Theentiresystemisdesignedtofunction completely offline, ensuring privacy, data security, and usabilityeveninareaswithlimitedornointernetaccess.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 12 Issue: 05 | May 2025 www.irjet.net p-ISSN: 2395-0072

6. CONCLUSIONS

TheResultAnalysisChatbotisapowerfulandeasy-to-use tool for managing academic performance information in schools.Byenablingautomaticextraction,processing,and analysisofresultdocuments,itsavestimeandminimizesthe possibility of human error. The inclusion of Natural LanguageProcessingenablesuserstocommunicatewiththe systeminaneasy,everydaymanner,andevenpeoplewith minimal technical proficiency can use it. Perhaps the greatest strength of this system is its offline mode, which ensuresdataconfidentialityandfacilitatesusageinremote or low-connectivity zones. By its modular design and unobtrusive coordination among elements like the text extractor,dataparser,Excelgenerator,andchatbothandler, thesystemprovidesprecise,efficient,andsecureacademic resultanalysis.Theprojectshowshowwebtechnologyand NLPintegrationcangeneratestrongeducationalresources thatarefunctionalandaccessible.

REFERENCES

[1] "Personal Chatbot For Documents"(2023) : Uday Bhamre, Harishankar Thakur, Naaz Sheikh, Chandan Patil, Prof. Bhaven Doshi, UG Student, Assistant Prof: InformationTechnology,TrinityCollegeofEngineering andResearch,Pune,Maharashtra,India

[2] "Implementing a Flask-based Chatbot for College EnquiriesusingSpacyandTensorFlow"(2023):Mr.P. Naresh,SamavedamVenkataramanaNagaPavan,Abdul RazzakhMohammed,ModepuTharun,NenavathChanti, Assistant Professor, UG Scholar, Department of InformationTechnology,VignanInstituteofTechnology and Science.R. Nicole, “Title of paper with only first wordcapitalized,”J.NameStand.Abbrev.,inpress.

[3] "CHAI: A CHatbot AI for Task-Oriented Dialogue with Offline Reinforcement Learning"(2022) : Siddharth Verma, Justin Fu, Mengjiao Yang, Sergey Levine, UC Berkeley

[4] "Multi-Purpose NLP Chatbot : Design, Methodology & Conclusion"(2023) : Shivom Agarwal, Shourya Mehra, PrithaMitra,QuantumDynamicsSAS,Marseille,France.

[5] "EmpoweringEducationwithAI:TheRiseofChatbotsin EducationalInstitutions"(2024):NomanAbdulRehman, Research Scholar, Swiss School of Management, Bellinzona,Switzerland,SyedMunassirHussain,College of Economics, Management & Information Systems, UniversityofNizwa,Nizwa,SultanateofOman.

[6] "EnhancingEducationalInteractions:AComprehensive ReviewofAIChatbotsinLearningEnvironments"(2024): Vipin Jain, Isha Singh, Madiha Syed, Sweety Mondal, DeepanshuRanjanPalai,SCSE,VITBhopal University, MP,India.ORCID:0000-0002-0099-3933

[7] "Educational Artificial Intelligent Chatbot: Teacher Assistant & Study Buddy"(2023) : Stergios Sozos, Dimitrios Zarris, Luleå University of Technology, DepartmentofComputerScience,ElectricalandSpace Engineering

[8] "Student Chatbot System: A Review on Educational Chatbot"(2023) : Sarthak Kesarwani, Titiksha, Sapna Juneja, Computer Science, KIET Group of Institutions, Ghaziabad,India

[9] "ApplicationofChatGPTinEducation:ACaseStudyand SWOTAnalysis"(2023):Boudadi,Siham,andMekkaoui, Samira. Education and Information Technologies. Springer.

[10] "Chatbots Applications in Education: A Systematic Review"(2021)ChineduWilfredOkonkwo,AbejideAdeIbijola, Department of Applied Information Systems, University of Johannesburg, South Africa. Computers andEducation:ArtificialIntelligence,Elsevier.