version 201601252034
 
 
Image courtesy of NASA
Feedback

Real Time Events

Social Media

News

Trends

Welcome to PeaceTech Lab’s (PTL) data portal for monitoring and reporting on online hate speech and offline violence for South Sudan. This platform, updated weekly, provides visualizations and analysis of social media hate speech data and offline incidents of violence and unrest, drawn from several sources, including Twitter, Facebook, ACLED, and GDELT. This portal has been created as a tool for various stakeholders working in the peacebuilding sector, particularly local organizations and activists focused on countering and preventing hate speech as well as those working to provide early warning and response (EWER) for violence and conflict in South Sudan.


Research in Kenya and several other countries, has demonstrated the potential for hate speech to instigate or inflame violence. For this reason, the monitoring and analysis of online hate speech should be included in EWER activities as a potential early warning indicator of violence.

Real-Time Events and Protests

GDELT ACLED
 
Legend

The above map displays recent violent events from two sources--which the user can switch between using the labels along the top of the map. The ACLED data, updated weekly, is a collection of violent events connected with ongoing conflict as well as public protests. The GDELT overlay shows hotspots where GDELT’s Event Database has captured trending news stories related to violent events. The time bar slider in the lower left corner can be used to view events from several days to several weeks in the past.

 

Hate Speech Volume

The line graph below shows the frequency of online hate speech posts and terms on a week-by-week basis. The data is drawn from social media monitoring platforms Crimson Hexagon and DataMinr. Content sources include Facebook, Twitter, Blogs, Forums, and Webpage Comment Sections.

 

The hate terms were identified through our Lexicon of Hate Speech process described above and on our website. The data range can be adjusted to focus on specific time periods. This visualization can be used to identify trends in the use of hate speech terms during specific periods, such as pinpointing when they experienced significant upticks.

 
Chart

Forecasts of Violence - Country Level

Chart

This graphic displays a time series of forecasted and actual weekly number of violent events in the country. The actual numbers are obtained from ACLED. The forecasted numbers are based on PeaceTech Lab’s statistical model that employs historical data from ACLED and GDELT and utilizes the social media hate speech data as a conditioning variable. The number to the right on the forecasted line indicates the number of violent events forecasted for next week. By clicking on an event category on the upper left, the data for that event category is dislayed. Use the timeline to adjust the time interval of interest.This data can be a useful tool in planning and prioritizing EWER and violence prevention activities.

 

Forecasts of Violence - State Level

 
 
 
 
 
     
The map on the left uses a heat gradient to illustrate the likelihood of violence against civilians in each municipalities for the upcoming week. These predictions are based on PeaceTech Lab’s statistical model that employs historical data from ACLED and GDELT and utilizes the social media hate speech data as a conditioning variable. The model puts out a probability of violence for each state (map on the left) and a probability threshold is employed to determine whether violence is expected or not (map on the right). Our model has an average overall accuracy of 86% with more than 1,000 predictions made. By clicking on an event category on the upper left, the data for that event category is dislayed. Use the timeline to adjust the time interval of interest. This data can be a useful tool in planning and prioritizing EWER and violence prevention activities.

Topic Wheel

The topic wheel below displays the top themes of conversation from public social media content including the hate speech term(s) selected. Use it to explore key topics (inner circle) and sub-topics (outer circle) in social media. Use the drop down menu to select one of the hate speech terms and a different date range to update the visualizations. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.

Top Sites

This chart shows which sites hate speech terms are most commonly used based on our monitoring. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.

Word Cloud

The word cloud below displays words most commonly used in social media posts including the hate speech term(s) selected. Use it to explore related words, themes, hashtags, and accounts in social media. Use the drop down menu to select one of the terms and a different date range to update the visualizations. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.

 

Content Sources

This chart shows on which sites hate speech terms are most commonly used, based on our monitoring. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, comments, Forums, YouTube.

GDELT News Analytics Word Cloud

From To
 
 
his live-updated wordcloud depicts top themes of news media related to hate speech in South Sudan. The data is generated by GDELT.

New Content Info

Introduction world cloud for News

Intro

Database

GLDELT

Configuration Portlet

DO NOT DELETE

this is done in order to avoid editing Liferay portlet code