blancaalcala/nlp-slack

W6 Project - Chat Sentiment Analysis Service

​ Description: You want to analyze the public chat messages (like slack public channels) of your team and create sentiment metrics of the different people on your team. The goal of this project is to analyze the conversations of your team to ensure they are happy 😃.

You will practice in this project:

• API (bottle)
• NLTK sentiment analysis
• Docker, Heroku and Cloud databases
• Recommender systems

Project Goals

​ Main goal: Analyze the conversations coming from a chat like slack ​

• (L1🧐) Write an API in bottle just to store chat messages in a database like mongodb or mysql.
• (L2🥳) Extract sentiment from chat messages and perform a report over a whole conversation
• (L3😎) Deploy the service with docker to heroku and store messages in a cloud database.
• (L4🤭) Recommend friends to a user based on the contents from chat documents using a recommender system with NLP analysis.
• (L5🔥) Do it real, use slack API to get messages and analyze the messages of our datamad1019 channel.
  • https://api.slack.com/ ​

TODO's - API Endpoints

​ You have to create an api with all this endpoints: ​

1. User endpoints

• (POST) /user/create
  • Purpose: Create a user and save into DB
  • Params: username the user name
  • Returns: user_id
• (GET) /user/<user_id>/recommend
  • Purpose: Recommend friend to this user based on chat contents
  • Returns: json array with top 3 similar users ​

2. Chat endpoints:

• (GET) /chat/create
  • Purpose: Create a conversation to load messages
  • Params: An array of users ids [user_id]
  • Returns: chat_id
• (GET) /chat/<chat_id>/adduser
  • Purpose: Add a user to a chat, this is optional just in case you want to add more users to a chat after it's creation.
  • Params: user_id
  • Returns: chat_id
• (POST) /chat/<chat_id>/addmessage
  • Purpose: Add a message to the conversation. Help: Before adding the chat message to the database, check that the incoming user is part of this chat id. If not, raise an exception.
  • Params:
    • chat_id: Chat to store message
    • user_id: the user that writes the message
    • text: Message text
  • Returns: message_id
• (GET) /chat/<chat_id>/list
  • Purpose: Get all messages from chat_id
  • Returns: json array with all messages from this chat_id
• (GET) /chat/<chat_id>/sentiment
  • Purpose: Analyze messages from chat_id. Use NLTK sentiment analysis package for this task
  • Returns: json with all sentiments from messages in the chat ​ ​

Links - API dev in python

• [https://bottlepy.org/docs/dev/]
• [https://www.getpostman.com/] ​

Links - NLP & Text Sentiment Analysis

• [https://www.nltk.org/]
• [https://towardsdatascience.com/basic-binary-sentiment-analysis-using-nltk-c94ba17ae386]
• [https://www.digitalocean.com/community/tutorials/how-to-perform-sentiment-analysis-in-python-3-using-the-natural-language-toolkit-nltk] ​

Links - Heroku & Docker & Cloud Databases

• [https://docs.docker.com/engine/reference/builder/]
• [https://runnable.com/docker/python/dockerize-your-python-application]
• [https://devcenter.heroku.com/articles/container-registry-and-runtime]
• [https://devcenter.heroku.com/categories/deploying-with-docker]
• Mongodb Atlas [https://www.mongodb.com/cloud/atlas]
• MySQL ClearDB [https://devcenter.heroku.com/articles/cleardb]