Data Science Capstone

Abstract

Research papers are one of the primary ways that researchers communicate their ideas to the rest of the scientific community. They offer a compact, familiar format to digest results, learn new methods, and review related work. Because of their important role in scientific communication, one’s research papers are often a deciding factor in whether one is accepted for an academic position, such as becoming a professor. In this project, we aim to better understand the relationship between the presentation of computer science (CS) research papers and their success. Analyzing this problem, however, involves many challenges, such as storing and analyzing data on millions of research papers, operationalizing paper success, and disentangling the effect of presentational factors from non-presentational factors. Previous work has analyzed the relationship between presentational factors and citation count, but they focus on a different field, whereas we focus on CS papers. For many of the presentational factors that we considered, there was an optimal range where the papers that were within that range had significantly higher citations than those outside of it.

Report

Feel free to check out the report here.