Dash

Documentation: https://dash.plotly.com/

Gallery: https://dash-gallery.plotly.host/Portal/

Streamlit

Documentation: https://docs.streamlit.io/en/stable/

Gallery: https://streamlit.io/gallery

· Dashboard is launched as a new tab in your browser as a “web app”

· The web app can be hosted locally on your computer or shared drive or can be uploaded to an online server (e.g. Amazon AWS, Google Collaborate, Microsoft Azure, etc.)

· Both allow for “debugging” on the fly being able to see the changes to the dashboard as changes are made in the code without having to relaunch the web app

· Integration with visualization packages such as Plotly and matplotlib

· They allow for rapid deployment of a dashboard with very minimal or no web design experience

· They allow for creating interactive elements that will filter and update your charts and DataFrames on the fly, such as dropdowns, radio buttons, sliders, checkboxes, buttons, etc.

· Creating a portfolio dashboard to view profits and losses, IRRs and current valuation metrics of investments with capability of filtering by sector, accounts, time periods and currency

· Creating a client invoices dashboard to view revenue generated by client, types of product, time periods, and geography

OS: https://docs.python.org/3/library/os.html

Pathlib: https://pathlib.readthedocs.io/en/pep428/

Shutil: https://docs.python.org/3/library/shutil.html

· Access and control to operating system files and folders

· Opening and closing files

· Copying, renaming, moving, and deleting files and folders

· Organizing hundreds of files and folders in an automated way

· Grabbing a list of all file names of one file type in a folder (e.g. all CSVs or PDFs)

· Cleaning up the names of multiple files

Website: https://python-pillow.org/

Documentation: https://pillow.readthedocs.io/en/stable/handbook/index.html

· Pillow is the friendlier, easier to use version of the PIL (Python Imaging Library) package in Python

· It provides image processing capabilities within Python and can handle and extensive list of image file formats such as PNG, JPEG, GIF, BMP, EPS and others

· Cropping, resizing and editing contents of images

· Open pictures of deal “tombstones” to be later analyzed and extract text with other OCR packages

· Import multiple logos of potential investment companies to be resized and cleaned up for a pitch presentation

Tabula

Documentation:
https://tabula-py.readthedocs.io/en/latest/

Camelot

Documentation:
https://camelot-py.readthedocs.io/en/master/

Comparison to Tabula:

https://github.com/camelot-dev/camelot/wiki/Comparison-with-other-PDF-Table-Extraction-libraries-and-tools

· Extract tables from PDFs

· Convert PDF tables into Pandas DataFrames

· Convert PDFs into CSV, JSON, HTML, Excel and other formats

· Dynamically find data based on keywords inside the tables

· Visualize the tables being extracted

· Finetune settings to account for tables with no borders and whitespaces between table cells

· Create data sets from multiple PDF files

· Extract, clean and convert tables from PDFs into Excel format in a more efficient and automated manner

· Extract key data from tables in SEC filings and financial reports

· Consolidate portfolio transactions from multiple PDF files from brokerage accounts

· Capture industry data from tables stored in PDF files

· Convert tables research reports into Excel files

Website: https://www.statsmodels.org/stable/index.html

Documentation: https://www.statsmodels.org/stable/gettingstarted.html

· Conducting statistical tests

· Classes and functions for estimation of many different statistical models

· Statistical data exploration

· Integration with Pandas DataFrames

· Linear regression to calculate beta for CAPM (Capital Asset Pricing Model)

· Time series analysis using ARIMA model

· Calculating betas of multiple factors in a portfolio using a multivariate linear regression model

Website: https://www.scipy.org/

Documentation: https://docs.scipy.org/doc/scipy/reference/

· Provides functions in mathematics, science and engineering

· Similar functions to other statistical programs such as MATLAB and Octave

· Functions built on top of the Numpy package

Website: https://scikit-learn.org/stable/

Documentation: https://scikit-learn.org/stable/getting_started.html

· One of the core machine learning packages in Python community

· Provides machine algorithms such as classification, regression, cluster detection, dimensionality reduction, data preprocessing and model selection

· Cleaning and preparing datasets for forecasting models: splitting data sets into testing vs training data, creating dummy variables for categorical fields, eliminating outliers

· Model evaluation: fine-tuning model parameters and analysing overfitting, comparing R-squared metrics and other model scores

· Forecasting more complex data that can’t be easily modeled using a linear regression model

· Categorizing data in an automatic fashion

· Determining credit rating of a company based on multiple independent variables, both numerical and categorical

· Finding the optimal capital structure and debt capacity of a company

· Determining the target price of a company using multiple key financial ratios and historical financials of a company

· Classifying customers of a company by spending habits to refine revenue buildup assumptions in an operating model

Website: https://opencv.org/

Documentation: https://docs.opencv.org/master/

· Open source computer vision and machine learning software library

· Used to open, process and transform images

· Used to identify special objects in pictures (e.g. eyes, faces, trees, etc.)

· Open multiple scanned images of legal documents

· Open and process logos of companies

· PyTesseract is the Python implementation of Google’s Tesseract technology

· Supports multiple image formats, including images processed from OpenCV or Pillow packages

· Supports multiple languages

· Extract all text from scanned purchase agreements

· Extract company names and financial figures from hundreds of deal “tombstones”

NLTK

Website: http://www.nltk.org/

Documentation: https://github.com/nltk/nltk/wiki

SpaCy

Website: https://spacy.io/

Documentation: https://spacy.io/usage

· Tokenization: Segmenting text into words, punctuation marks,

· Part of Speech (POS) Tagging: Assigning word types to tokens, e.g. verb or noun

· Named Entity Recognition (NER): Labelling named “real world” people, companies and locations

· Text Classification: Assigning categories or labels to a whole document, or parts of a document

· Both packages have models of “taught” words that act as starting dictionaries

· Extracting key words from press releases, essays, and text documents

· Translating words from one language to another

· Analyzing sentiment of articles

· Extracting key information from SEC filings

· Summarizing a company’s quarterly earnings press release seconds after it is filed

· Extracting all companies mentioned, dates and financial figures from hundreds of articles on an industry website