This repository serves as a brief overview of my stage gate one project for CSCI 494-007, National Security Data Science in Action. To demonstrate my competency using Python and related data analysis libraries, I implemented an off-the-shelf neural network using TensorFlow and Keras. Given my inexperience with these two high-level libraries, I followed the "Basic Text Classification" developer guide from the official TensorFlow documentation. This project involved training a simple feed-forward neural network to perform sentiment analysis on film reviews from the IMDb database. The first task in this guide involved collecting and preprocessing the underlying data. Though I originally planned on using Pandas for this (a library that I have used in previous projects for such things as data wrangling), I discovered from the guide that using Pandas was unnecessary in this case; the TensorFlow library ships with a comprehensive repository of data preprocessing utilities (see my Jupyer notebook for a more in-depth analysis of these utilities). After processing these data, I followed the guide in defining a sequential model using the Kears API (once again, see my Jupyter notebook for an in-depth analysis of the model’s configuration), and trained said model using a binary cross entropy loss function. Following this, I analyzed the loss and accuracy of the model, using matplotlib to create some simple visualizations. After reporting on these results, I decided to extend the scope of this project by applying my learning to implement an early stopping callback. The goal with this callback method was to address the overfitting issue that presented itself in my (slightly contrived) model. Although I have much more to learn in the way of these tools, I feel as though I have satisfied the goal that I set out to accomplish: gaining experience working with high-level APIs for solving data science problems.