Automated CX Detection through Social Networks - Scripts & Dataset

Published: 21 July 2021| Version 1 | DOI: 10.17632/wt99brrp7f.1
Contributors:
Leonardo Kuffo,

Description

This dataset is comprised of the following: a. Data Extraction scripts (DataExtraction directory): Scripts to extract data from Twitter, Foursquare and Facebook using the respective platforms APIs. You must use Python 3.x to run the script called CX_EXTRACTOR.py to extract data for any of the three platforms. You must also provide your own API KEYS for each platform. b. Classified messages from Twitter and Facebook into one of our detected CX dimensions: Product Experience, Customer Service, Digital Experience, In-Store Customer Experience and Events Experience (Data directory). Each dataset is in a csv format and contains the following columns: unique identifier, our classifier predicted class, authored date, establishment brand, message text, sentiment score. c. Trained models binaries & scripts to test the models using a toy dataset (Classification directory). You must use Python 3.x to run the script called model_testing.py. You can input your own test data inside the test.csv. After executing the script, results will be written into prediction.csv file. Programming Language: Python 3.x

Files