The current version generates estimated current ratios using Twitter tweets.
You can find specific instruction in the JupyterNotebook
Build Financial Analysis From Scratch-Deliver.ipynb
- Required packages are located in the
requirements
file. You can install the first three packages,scrapy
mongodb
andmysql-connector
by running the cell in the above notebook.
The files in the folders are described as follows:
Build Financial Analysis From Scratch-Deliver.ipynb
-- Interactive Jupyter Notebook to make predictionStructured.py
-- supporting functions to construct structured features for predictionParseTweet.py
-- supporting functions to parse tweets to unstructured features for predictionSelectInput.py
-- generate ipywidget to take inputcountry_dict.pickle
-- translate country name to two letter ISO codeus_state_list.pickle
-- list of two letter US state namescanada_state_list.pickle
-- list of two letter Canada state namessic_reg.pickle
-- translate division and industry to four digit sic codedatax_header.pickle
-- regulate structured dataridgemodel.pickle
-- first part of the pre-trained model using mean estimatorsentimentmodel.pickle
-- second part of the pre-trained model using sentiment analysis
The following files are included to support TweeterScraper App:
TweeterScraperReadMe.md
, no need to read and do not need to follow its instructionLICENSE
, license of the appscrapy.cfg
and folderTweetScraper
are part of the App. They should not be removed.- Temporary folder
Data
and temporary fileget_tweet.sh
will be generated to collect tweets