###Install Elasticsearch: https://www.elastic.co/guide/en/elasticsearch/reference/current/_installation.html
Allow dynamic scripting:
in elasticsearch-x.y.z/config/elasticsearch.yml, add:
script.inline: on
script.indexed: on

###Install Twython, Elasticsearch modules: pip install twython
pip install elasticsearch

###Set configs: Rename file "config.py.example" to "config.py". You can now set collect parameters.

###Get Twitter access tokens: Head over to https://dev.twitter.com/apps and register an application. After you register, grab your applications consumer_key, consumer_secret, oauth_token, and oauth_token_secret.
Paste them in the config.py file.
If you have several applications registered on twitter it is even better: paste all your access tokens in the config.py file. Quick_twython is designed to use all available access tokens to retrieve tweets.

You can have access to my test word2vec model trained on 2 million french tweets collected between October 3rd and October 7th 2016. Tweets were treated to remove urls, "@" mentions, punctuation and uppercases. Gensim phrase detection tool was used to detect common bigrams such as "françois_hollande". The model is available at /quick_twython/word2vec_models/