SentiMap

Web app for realtime sentiment analysis and mapping of tweets.

##Initial Configuration## This application requires both Redis and MongoDB. In order to run it you'll need to download each and then start both mongod and redis-server.

You'll also need to configure both python files, and the JavaScript file plot.js.

###Stream_Mod.py### First install all dependencies:

pip install nltk numpy sklearn gensim pymongo setproctitle

Then download the NLTK stop words file as described in the documentation.

>>> import nltk
>>> nltk.download()

Next you'll need to sign up for Twitter API access credentials by registering for an application, then filling in your Twitter Oauth credentials in the variables:

access_token = "ACCESS_TOKEN"
access_token_secret = "ACCESS_TOKEN_SECRET"
consumer_key = "CONSUMER_KEY"
consumer_secret = "CONSUMER_SECRET"

The same is necessary for the Bing Translator API. Head over to the Azure Marketplace and register for authorization credentials, then subscribe to the Translator dat a product if you want to access the translation features of the streamer so that you can handle multiple languages. (NOTE: The Bing Translator API only allows for 2,000,000 characters per month for the free tier.)

When you are finished you should supply your access credentials via the variables:

client_ID = "CLIENT_ID"
client_secret="CLIENT_SECRET"

Finally you'll need to download the Google Word2Vec Training Vectors, extract the binary file, and place it in static/Data/.

Run Stream_Mod.py to initialize the database and start classifying and storing tweets.

###SentiMap.py###

Install dependencies:

pip install redis flask pymongo

Start SentiMap.py to run the flask server. The web server will run by default on localhost:5000.

###Runtime### When you want to start, first run Stream_Mod.py to set up the database structure and start listening for tweets (this will take a while to start, because the Google vectors binary file is large and takes a while to load into memory). After you've started Stream_Mod.py you can move on to starting SentiMap.py.