Project : NLP Sentiment Analysis
Using TwitterAPI and tweepy, extract tweets for sentiment analysis. "How do pepople feel about traveling during COVID19?" Tweets were extracted on July 2020. Analyse current customers' trend of 3 different locations (Melbourne, Berlin and New York). sentiment_analysis demonstrates in more details.
- Data extraction
- Exploring and cleaning data
- Feature engneering
- Building models
- Analysing sentiment
Motivation
The study of tourist behaviour is crucial for travel industory after COVID19. I believe that forcasting customer trend and demand will have a great advantage for a business in tourism.
Files' and folders' descriptions
Folders
- data Folder containing data files
- image Images used for sentiment_analysis
- models Models saved for pickle
- utils Utils folder
- data_cleaning_utils.py for cleaning tweets data, used in data_cleaning.py
- data_extraction_utils.py for extracting tweets, used in api_data_retrieval.py
Files
- sentiment_analysis Analysing sentiments and model outcome by using graphs and images
- api_data_retrieval.py Extract tweets (No location specify, it is for training datasets)
- data_cleaning.py Cleaning train dataset using utils
- feature_engneering.py Feature engneering: creating TF-IDF and normalized numeric variables