Anustha Shrestha

I am an aspiring statistical analyst with prior experience in Financial Planning and Analysis in the Insurance Industry. I combine analytical skills, technical expertise and business knowledge to present clear and concise reporting for non-quantitative audiences.

In my three years of experience as a finance analyst, I saw how data has shaped the changing landscape of finance and I wanted to be a part of the change. Therefore, I joined graduate program in Statistics at Baruch College to get a deeper understanding of statistical concepts and programming languages. I am sharing some of the projects that I have taken on in the past few years and hope to learn more about statistics and data science.

Feel free to contact me at anusthas@gmail.com or view my LinkedIn Profile

Projects

Machine Learning

Crime and Communities: Classification Analysis of High and Low Crime Rates - Evaluated performance of LASSO, Ridge, Random Forest, SVM - in classifying 2000 U.S. communities into “high” or “low” crime areas. Randomly sampled 100 train-test splits, compiled error rates for each iteration, and created box plots for comparison. Optimized hyperparameters for LASSO, Ridge and SVM. Given complexity, Ridge performed the best (17% error rate).

Time Series

Time Series Modeling and Intervention Analysis on Tourist Arrivals in Nepal - Predicted 2019 and 2020 monthly tourist arrivals in Nepal using a multiplicative seasonal ARIMA model with data from 1992 to 2018. Intervention analysis identified 2015 earthquake as temporary change outlier with an immediate effect of 64% reduction of arrivals in the first period and subsequent gradual decrease of impact, which was non-existent by the end of 2018.

Shiny Dashboard

High School and Beyond Survey Dashboard - National Center of Educaitonal Statistics’ High School and Beyond Survey was used to create a dashboard for exploratory data analysis using various univariate and bivariate plots and simple linear regression model.