In this project we had worked for Covid19 Twitter Sentiment Analysis.
The dataset which we used to train our data consisted of tweet along with the original user sentiment at time of tweet. From a survey it also consisted of time taken for writing tweet and 1-10 scale for each individual feeling used
- Knowledge of Django (This project used Django as framework)
- Knowledge of RNN (Bidirectional LSTM) Algorithm
- Knowledge of NLP
- Twitter developer id (for live sentiment analysis)
- Knowledge of IBM Cloud (for deployment)
Dataset link : https://arxiv.org/pdf/2004.04225.pdf
conda install -c anaconda pillow
conda install -c conda-forge matplotlib
conda install -c anaconda seaborn
conda install tensorflow
python manage.py runserver
~ using function
~ connections made and call in url.py
~ functions in views.py
~ url activate in html file
html(activation of function)-> url(cheking function) -> views.py(checking function definition)
Main->Static Folder-> Trial Analysis2
Main-> Templates->Html
Main-> Twwet_Dashboard-> models.py->(forms)
urls.py->(paths for webpages)
views.py-> (working functions of backend)
Management-> commands-> bluemix_init.py(for hosting)
Jupyter_notebooks->Multifeeling_value.ipynb( For multiple sentences/lines output)
->twwet_me.ipynb ( For single line output )
tokenizer-> tokenizer_SAVED_OOV.pickle (tokenize (oov token))
(padding fn)[in jupyter notebook]
dataset->hell.csv
weights-> multi_traget_feeling.hdf5 (multiple feelings for a particular text as outcome as bar graph in dashboard)
-> first_model_feeling1longonly_NEWMAIN.hdf5(single feeling as a summary of multiline text)
Documentation link : https://drive.google.com/file/d/1E4MIv14svusdBCJkg6E3bNFCsnhkysSj/view?usp=sharing
PPT Link: https://drive.google.com/file/d/1fyJgoPZ6R57VXBwPeVX8mYqay2lOtYBV/view?usp=sharing
Video Link : https://drive.google.com/file/d/1LYOSQZQHyf8ZVZgsoek9iCeLQJCBde99/view?usp=sharing
Sample text to be test: https://drive.google.com/file/d/1D_1HkI-xMGVw1PotbrnCsSswF8Byvbis/view?usp=sharing
All Files link : https://drive.google.com/drive/folders/1PzMCkXa3VQy1cj36E2ulXMNXDTC2R6jk?usp=sharing
- http:https://www.cs.columbia.edu/~julia/papers/Agarwaletal11.pdf
- https://towardsdatascience.com/twitter-sentiment-analysis-based-on-news-topics-during-covid-19-c3d738005b55
- https://arxiv.org/pdf/2003.05004
- https://www.jmir.org/2020/4/e19016
- https://arxiv.org/pdf/2003.10359
- https://www.researchgate.net/profile/Kia_Jahanbin2/publication/339770709_Using_twitter_and_web_news_mining_to_predict_COVID-19_outbreak/links/5e84d4db4585150839b508b7/Using-twitter-and-web-news-mining-to-predict-COVID-19-outbreak.pdf
- https://arxiv.org/pdf/2003.12309
- https://arxiv.org/pdf/2004.04225
- https://towardsdatascience.com/how-are-americans-reacting-to-covid-19-700eb4d5b597?source=rss----7f60cf5620c9---4