sjdonado / sentime

Citizens sentiment analysis through their tweets

Sentime

Datasets

2020_04_27_dataset.csv: 30633 tweets:

• 9959 Positive
• 6123 Negative

2020_05_24_dataset.csv: 53366 tweets:

• 19534 Neutral
• 17483 Positive
• 12376 Negative
• 3972 Mixed

How to run?

• Server Go to server folder and run:

  docker-compose up

• Client Go to client folder and run:

  npm start

Model preprocessing

Go to model folder and follow the next steps:

1. Setup

• Install virtualenv pip install virtualenv
• Create new virtual env virtualenv . and activate it source bin/activate
• Install dependencies pip install emoji

2. Run model/scraper.py

  # Example
  python3 model/scraper.py

3. Run model/preprocess.py ${TODAY_DATE}

  # Example
  python3 model/preprocess.py 2020_04_27

4. Send tweets_parsed to AWS Comprehend, download the output and rename it as aws_output, finally move it to model/data/${TODAY_DATE} folder
5. Run model/generate_dataset.py ${TODAY_DATE}

  # Example
  python3 model/generate_dataset.py 2020_04_27

6. Create a folder called Sentime in Google Drive and open sentime.ipynb in Colab
7. Upload the generated datasets to sentime/${TODAY_DATE}/train/data.csv and sentime/${TODAY_DATE}/test/data.csv

Connect with database using pgAdmin

1. Go to http://localhost:5431
2. Login with email: [e-mail] and password: root_12345
3. Create a new server
4. Fill the form as follows:

• Use dbinstead of localhost in the address field
• User: sentime_user
• Password: root_12345

Load model

Go to http://localhost:5000/search/status

Test predictions

Go to http://localhost:5000/search/test?text=Juan%20est%C3%A1%20tirando%20c%C3%B3digo%20con%20cule%20de%20felicidad

Signup link

Go to http://localhost:5000/users/signup/4f3a0ca08e906