Fix Telecom Customer Churn Prediction Example Notebook #742

Three small issues with the telco churn walkthrough:

1. The docker container used for loading the example notebooks has all of the packages for the "basic" notebook environment, but not for the "telco" notebook. (eg, pillow, python-slugify, etc.)
2. The environmental variables do not need to be set, as they are part of the config in the the docker-compose. The environmental variables set to localhost, but in the config are core. Using the code to set the environmental variables in the notebook results in an error.
3. For making predictions, there is an entity_df variable that is referenced in the telco notebook - but that variable is not instantiated anywhere.