Sandesh Shrestha

Minnesota State University Mankato
|2018-2022|
Statistics
Minor: Computer Science & Mathematics
|2024-2025|
Data Science
Projects
Reader's Graph
  • Developed a user-friendly website that empowers book enthusiasts to share their favorite content and engage with a diverse community of readers.
  • Utilized advanced technology stacks by constructing the backend infrastructure with Django and Python, while creating a responsive frontend using Next.js and NextUI for a seamless user experience.
  • Demonstrated expertise in REST API integration, harnessing the power of Axios to make efficient calls to backend routes and proficiently implementing dynamic routes within the frontend.
  • Leveraged AWS S3 bucket to securely store images uploaded by users, ensuring data integrity and reliability.
  • Employed AWS Lambda functions to enhance user experience by generating image thumbnails before storing them in the S3 bucket, optimizing website performance.
  • Enhanced website efficiency by implementing Amazon CloudFront for indirect access to images stored in the S3 bucket, guaranteeing swift content delivery to users.
  • Seamlessly integrated the Google Books API to enable readers to effortlessly search and share their favorite book quotes or lines, enriching the platform with a vast array of literary content.
Loan Approval Predictor
  • Employed Machine Learning techniques to develop predictive models for assessing loan eligibility based on a range of socio-economic variables, including gender, marital status, dependents, and educational background.
  • Leveraged data science libraries such as scikit-learn and matplotlib to construct robust machine learning models.
  • Skillfully addressed data anomalies, including the identification and treatment of outliers, while ensuring data consistency through normalization procedures and visualization tools like heatmaps to evaluate variable correlations.
  • Successfully engineered and fine-tuned multiple machine learning models, encompassing Logistic Regression, K-Nearest Neighbors (KNN) Classifier, Multi-Layer Perceptron (MLP) Classifier, Gaussian Naïve Bayes, and RandomForest Classifier.
  • Employed the GridSearchCV technique to optimize model parameters, culminating in an impressive model accuracy rate of 84%.
Song Lyrics
  • Developed a user-friendly website that efficiently displays song lyrics by seamlessly collecting song and artist details from users.
  • Integrated a music API to dynamically fetch and provide lyrics to users, enhancing the website's functionality and user experience.
  • Employed Django framework and Python for the robust backend development, ensuring seamless data processing and storage.
  • Designed and implemented the frontend using HTML, CSS, and JavaScript, delivering an aesthetically pleasing and responsive user interface.
Breast Cancer Prediction
  • Developed a predictive model using Machine Learning techniques to assess the probability of Breast Cancer based on an analysis of malignant tissue attributes, including size, shape, and texture.
  • Applied Supervised Machine Learning with a focus on utilizing the Decision Tree Classifier algorithm to construct and fine-tune the model.
  • Employed data visualization tools such as Matplotlib and heatmaps to create graphical representations of the dataset, highlighting variable correlations.
  • Conducted pruning procedures to eliminate less relevant branches in the decision tree, thereby enhancing model efficiency.
  • Implemented the concept of CCP alpha to prevent overfitting, ensuring that the model maintains a high accuracy rate, achieving a remarkable accuracy level of 97%.
MUDAC
  • Collaborated in the Midwest Undergraduate Data Analytics Competition (MUDAC) hackathon in 2019 as part of a five-member team representing Minnesota State University (MNSU).
  • Spearheaded research focused on assessing the impact of climate change on water resource pollution within the region of Minnesota State.
  • Conducted in-depth investigations into historical occurrences and significant events related to water resource alterations brought about by both natural and human-induced factors.
  • Utilized the findings from our research to contribute to the development of a Machine Learning model, providing valuable insights and data-driven solutions.