Messy Code for an application to visualize geosocial trajectories needs a lot of refactoring removing unused stuff etc...
It preprocesses tweets ( language and sentiment detection ), stores that into a postgre DB. A web application uses the data and calculates trajectories for a given clustering strategy.
Local Php Server, Postgres DB (used version 9.4.1) with Postgis extension
contains the Java code for preprocessing tweetdata and storing into a Postgres DB.
contains result of training data for language, topic and sentiment classification
contains the web application that is build on top of cesium.js and uses d3.js for data manipulation and processing
Tweetdata is in CSV model see example.csv for structure
http://stanfordnlp.github.io/CoreNLP/ https://sourceforge.net/projects/simmetrics/ http://www.csvreader.com/java_csv.php http://alias-i.com/lingpipe/