About Me

Sapna Sharma

A Data Science Enthusiast

My Career

Harrisburg University

PhD - Data Science
Jan-2025 - present

CVS Health

Data Engineer
March-2022 - present

auxi.ai

Machine Learning Intern
Working on LayoutGANs to to automate slide formatting and manual corrections in business presentations (2021-02-01 - 2021-05-01 )

Northeastern University

Teaching Assistant
DS2000 An introduction to programming for data scientists (01-09-2020 - 31-12-2020)

Fellowship.AI

Machine Learning Fellow
Deployed end to end Wound Tissue Classification| Developed pipeline for Speaker Diarization by video and audio of a zoom meeting (01-05-2020 to 30-08-2020 )

Northeastern University

Teaching Assistant
DS2000 Programming with Data (01-01-2020 to 30-04-2020)

Northeastern University

Data Analyst

Radio Network Optimizer,BSNL

Data Analyst

Web Developer

www.losefat.co.in (Developed in Wordpress)

Web Developer

My Skills

My Projects

Capacity Planning and Traffic Forecasting for Boston's Bluebikes

● Demand Calculation and Capacity Planning ● Time series Traffic forecasting

Built Data Warehouse for Pubmed Data

● Built relational database ,ERD and Snowflake Schema ● Performed Data Mining and data visualizations using MySQL and R

Artistic Style Transfer with Neural Nets

● Implemented the paper "A Neural Algorithm for Artistic Style" ● Use pre-trained VGG-19 model for reconstructing the style and content image

Credit Card Fraud Detection

● Explored models like Decision Trees, Logistic Regression, SVM, XGBoost, Random Forest, ANN ● XGBoost Classifier performed best with a Precision of 0.36, Recall of 0.78, f1-score of 0.49

Airbnb Analysis using Causal Modelling in Machine Learning

● Applied Causal Inference model with pyro to calculate the return of investment of hosts that list their properties on Airbnb on the basis of location , number of bedrooms, number of bathrooms and mobility score

House Price Prediction

● Performed exploratory data analysis, pre-processing and feature engineering for more than 70 features for model building ● Tuned hyperparameters of XGBoost regressor and achieved an RMSE score of 13.439%

Meeting Topic Modelling (LDA)

● Meetings’ topic modelling (LDA)● Performed Data Cleaning,Data transformation,Hyperparameter Tuning,Visualize Results,Testing on Unseen Document

Image Classification (fastai)

● A project to classify wound tissue. ● Wound Tissue Analysis (Web Scraping and training models using fastai library and resnet models)

Image Visualization

A project to do Image Visualization using t-SNE dimentionality reduction using PCA and Image segmentation with K-means clustering, GMM, Hierarchical clustering.