In an opportunity from General Assembly and Track Maven, we gather data-driven insights to help brands drive Twitter Engagement.
This is a github repo of my contributions to a group project.
You can find the presentation we gave here: https://docs.google.com/presentation/d/178vgdcrY4oBZ1QXHJYtTjZ4Be9C6k4SSbnmFLfxNoB4/edit?usp=sharing
My contributions include:
- Using AWS Rekognition to turn our image data into text data
- Apply NLP techniques to analyze the text in our image tags
- Performing Latent Dirichlet Analysis on the image data to break the images into topic categories
- Exploring how the addition of image data affected the model constructed by Dale and Diego
- Creating Decision Trees for each brand - to provide a tool for users to apply human judgement and our data insights to their content creation
Other group members were:
Dale Wahl - https://github.com/dale-wahl
Diego Rodriguez - https://rodriguezda.github.io/
Kyle Santana - http://kylesantana.com
Matt Bollinger - https://github.com/mlybollinger