Civil Engineering • Structural Analysis
Performance Based Seismic Design
Developed performance-based seismic design framework for RC structures using pushover analysis and FEMA guidelines for multi-hazard resilience assessment.
Posts by Collection
portfolio
Machine Learning • Audio Processing
Voice Gender Recognition
Developed audio classification system for voice gender recognition achieving 96% accuracy using MFCC features and SVM/Random Forest classifiers.
Web Development • Drupal
Website Development - Indian Red Cross Society
Designed and deployed Drupal-based CMS for volunteer registry and donor database, reducing manual data entry by 50% and supporting 10,000+ monthly visitors.
Machine Learning • Classification
Adult Income Classification
Built ensemble classification models on UCI Adult Census dataset achieving 87% accuracy in predicting income levels using gradient boosting and feature engineering.
Environmental Science • Impact Assessment
Taj Mahal Environmental Impact Assessment
Comprehensive environmental impact assessment analyzing air quality, water pollution, and structural degradation threats to the UNESCO World Heritage monument.
Time Series • Forecasting
Statistical Analysis and Forecasting of Solar Energy Production
Developed time series forecasting models (ARIMA, Prophet) for solar energy production achieving MAPE of 8.5% on 3-year historical data.
Statistics • DOE
Design of Experiments (DOE) Analysis
Applied factorial design and response surface methodology to optimize industrial processes, achieving 25% improvement in process efficiency.
Deep Learning • CNN
Convolutional Neural Networks for Image Classification
Implemented CNN architectures from scratch achieving 92% accuracy on CIFAR-10, exploring various optimization techniques and regularization methods.
NLP • Text Analysis
Automated Plagiarism Detector
Developed multi-algorithm plagiarism detection system using TF-IDF, cosine similarity, and Jaccard index achieving 91% detection accuracy.
Machine Learning • Active Learning
Active Learning for Forest Cover Classification
Developed active learning framework achieving 94% accuracy with 30% less labeled data using uncertainty sampling and query-by-committee strategies.
Policy Analysis • Environmental Science
Madhav Gadgil – Kasturi Rangan Reports Analysis
Comprehensive policy analysis comparing two landmark environmental committee reports on Western Ghats conservation and their implementation outcomes.
Machine Learning • IoT
Applications of ML/IoT in Concrete Technology
Published research on ML/IoT integration in concrete strength prediction and quality monitoring, achieving 89% accuracy with Random Forest and XGBoost models.
Environmental Science • Regression
Air Pollution Abatement and Modeling
Developed multivariate regression models predicting ambient air quality (PM2.5, NO2, O3) using meteorological, traffic, and industrial emission features with 85% accuracy.
Deep Learning • CNN
LEXNet for Internet Traffic Classification
Implemented CNN architecture achieving 98.7% accuracy on ISCXVPN2016 dataset for encrypted traffic classification and VPN detection.
Causal Inference • Sports Analytics
English Premier League Spending Analysis
Quasi-experimental analysis examining causal effects of squad investment on team performance using synthetic control and difference-in-differences methods.
NLP • Deep Learning
EV Sentiment Analysis Using Transformer Models
Analyzed 1.1M+ Reddit posts on electric vehicles using transformer-based NLP models, achieving 91.6% accuracy in sentiment classification. Presented at AAPOR 2025.
Machine Learning • Sports Analytics
Player Market Value Analysis in Elite European Football
Econometric analysis of transfer market valuation for 600+ players across Europe’s top 5 leagues using regularized regression and feature engineering.
publications
Research publication
Achieving Sustainability in Supply Chain during Disruption Times: Role of Industry 4.0
Designed and executed a qualitative survey study with 200+ automotive supply chain experts, employing fuzzy-set AHP and ISM-MICMAC analysis to rank Industry 4.0 technologies and green supply chain practices.
Shrivastava, N., et al. (2024). "Achieving Sustainability in Supply Chain during Disruption Times: Role of Industry 4.0." Advances in Data-Driven Computing and Intelligent Systems. Springer.
Research publication
Analyzing Public Perceptions and Sentiment of Electric Vehicles in the Era of Sustainable Transformation
Engineered a multi-source sentiment analysis pipeline processing 1.1M+ social media posts from Reddit and NYT, achieving 91.6% classification accuracy using DistilBERT and comparing 10 LLM variants for sentiment bias.
Chakravarty, S., & Shrivastava, N. (2025). "Analyzing Public Perceptions and Sentiment of Electric Vehicles in the Era of Sustainable Transformation." 80th Annual AAPOR Conference. St. Louis, MO.
talks
Analyzing Public Perceptions and Sentiment of Electric Vehicles in the Era of Sustainable Transformation
Presented transformer-based EV sentiment research at AAPOR, covering a 1.1M+ post pipeline, 91.6% DistilBERT accuracy, and measurable LLM sentiment bias.
teaching
Graduate Assistant
Canvas LMS Infrastructure & Course Administration
Graduate assistant role focused on Canvas LMS redesign, reusable course infrastructure, and workflow support across JPSM instruction.
Graduate Course - Teaching Assistant
SURV735: Data Privacy and Confidentiality
Teaching assistantship for SURV735 supporting graduate instruction in data privacy, confidentiality, and privacy-preserving statistical methods.