Case Example - Twitter U.S. Airline Sentiment Analysis

This case example was done within the scope of the seminar paper "Social Media Analytics" for the 2018/2019 Data Warehouse Systems Seminar (703615 SE/2 SE) by Univ.-Prof. Mag. Dr. Maier Ronald

The case example was developed using python and jupyter along with several python libraries that can be found in the requirements.txt file. The main analysis can be found in the twitter-airline-sentiment-analysis.ipynb notebook.

Update 21-01-2019

The twitter-airline-sentiment-analysis-vader-comparison.ipynb notebook was included to show a comparison of human verified sentiment classification to VADER classification.

Authors

• Johannes Ebster, 3bst0r
• Daniel Egger, dnlggr
• Jannik Siebert, janniks

Development

Default python3 and pip are required to run this notebook.

Set up a .venv virtual environment

python3 -m venv .venv

Activate the .venv virtual environment

. .venv/bin/activate

Install the requirements defined in requirements.txt

pip install -r requirements.txt

Start jupyter and open twitter-airline-sentiment-analysis.ipynb

jupyter notebook

Attributions

• Twitter US Airline Sentiment - Analyze how travelers in February 2015 expressed their feelings on Twitter
• Based on an article from the data scientist Martín Pellarolo
• Crowdflower