ICWSM_Emoji ❤️💜💔💨🙏
This repository contains the dataset for the ICWSM emoji paper titled I stand with you: Using Emojis to study solidarity in crisis events published at ICWSM Emoji workshop, 2018, Stanford, USA.
Two large datasets made available
Due to Twitter's restrictions, we are only able to share the tweet ID's
1. **Paris Novemer Attack** - Tweet Id's are available in the paris_dataset folder
2. **Irma Hurricane Dataset** - The files 06.txt to 12.txt contain the tweets ids of the english tweets of hurricane irma collected between Sept 6th 2017 to Sept 12 2017.
This repository contains a mixture of codebase written in R (predominantly) and Python.
Please download the tweets with the help of the tweet IDs using this repository in Python or use the rtweet library for R.
**Parsing JSON dataset**
parsing_json.Rmd -> Extracting tweets and necessrary fields from JSON.
Depends on the annotation made available in the annotations folder to further pre-processs the dataset.
Classifier
- Prepare the excel file with tweets and label.
- Run the classifier.
Steps needed to produce the figures in the paper
-
Prepare the excel file with required fields
text (tweet), labels(-1 or 1), postedtime, locations, latitude, longitude, count (count of emojis in a tweet), description(type of emoji), country, community (affected or not), estTime (converted time), day and hour -
Producing Time Series - this files gives an overview of how to produce the file that is needed for counting emojis and creating a time series file
-
Geo-tagging sample code is made availble for the IRMA dataset. This helps give the location mentioned in the tweets with a latitude, longitude and country.
Reproducing Co-occurence Network Run the irma_tweets_emoji_analysis.Rmd and paris_tweets_emoji_analysis.Rmd with the prepared dataset.
Should produce these figures
Reproducing Diffusion Plots Run the irma_diffusion.Rmd and paris_diffusion.Rmd with the prepared dataset.
Should produce these figures - Diffusion Across Days
Diffusion based on location
To Do:
-
Add Classifier - Add requirements
- Add classifier parameters
Please contact ssantha1@uncc.edu if you have any further questions with regards to this project.