This README documents necessary steps to get the application up and running.
- Description: This is the final release of the version 1.1 of Tweet Relevance System. This application fetches trending topics from the twitter and tweets for them. The fetched tweets are filtered to find tweets of popular people/celebrities. The database of celebrity twitter accounts is maintained in the mongodb. The end result is a list of trends with each containing a list of tweets by popular people/celebs.
This release includes classification of Twitter trends/hashtags into categories. The classification is done by using a Naive Bayes model trained on the categorized data. The model is stored in the mongodb.
- Version: 1.1
- Samana Katti (sk6021@g.rit.edu)
- Pankaj Vasant Uchil (pu3876@g.rit.edu)
- Kantha Girish Gangadhara (kg2605@g.rit.edu)
-
The code is written in python-3.5. The below mentioned dependencies are required to be installed
-
Configuration : tweets/config.ini
[twitter keys] The keys are required, can be obtained by creating application on apps.twitter.com
[training] Specify the name of folder containing traning data and name of stop words file
[log] Specify the location and file names for logging script results.
[mongodb] Specify the name of database to use for the application -
Dependencies :
python 3.5 or higher,
python-twitter (https://pypi.python.org/pypi/python-twitter/)
pymongo
django version 1.1
Please make sure that all the dependencies mentioned above are installed and all the files and paths required are specified in config.ini.
- Training classifier: - python3 train_classifiers.py
- Populate application database with tweets: - python3 populate_feed.py
- Start web server: - python3 manage.py runserver