Skip to content

Latest commit

 

History

History
104 lines (68 loc) · 5.54 KB

README.md

File metadata and controls

104 lines (68 loc) · 5.54 KB
Raw

StockPerformanceIndicators

This projects main functionality is the visualization of the following stock performance indicators (called "stock PI's" in the following):

  1. Piotroski F-Score
  2. PEG-ratio

Table of Contents

  1. Requirements

  2. Project Description

    2.1 User Interaction

    2.2 Stock PI Cache

    2.3 CSV Downloader

  3. Structure

  4. Known Issues and Troubleshooting

  5. Sample Stock Tickers

Requirements

  • Python 3 or greater
  • Installed packages listed in requirements.txt (e.g. with pip install -r requirements.txt)

Project Description

User Interaction

After running the main.py, the user is prompted to:

  1. Select a stock PI for visualization
  2. OR: Download stock PI's from Yahoo! Finance into a CSV (the file has to be filled with stock tickers and indices beforehand and has to have a "Country"-column - see example.csv

To visualize a stock PI, the user has to select a main stock ticker used for comparison to the same PI of other stocks. The user can either use all instances of the stock PI in the cache/ directory for comparison or specify the stocks used for comparison with the following filters:

When visualizing the PEG-ratio, the user additionally has the ability to select the maximum range of the ratios to be displayed in the plot. If the PEG-ratio of the main stock is not in this range, the user is alerted and can specify a new range.

After all user-inputs, one of the following plots is displayed:

PEG-ratios Piotroski F-Scores

The red data-point marks the PI of the main stock specified by the user and it's ticker is constantly displayed. When hovering over a black comparison-data-point it's ticker is also displayed. The blue line indicates the mean of the specified PI's and the light-blue area indicates their standard deviation.

Stock PI Cache

When comparing stock PI's, one thing is crucial: A large database. To retreive this data I used the libary yfinance which scrapes Yahoo! Finance. Unfortunately the computation time for the requests being send to Yahoo! Finance is very long. This results in a huge detriment to user experience if real-time scraping of a large stock-database is performed.

My solution for this problem: I implemented data_into_csv.py to pre-download stock PI's and save them in the CSV files located in cache/.

The stock PI's are always selected out of cache/. However, if cache/ does not contain the main stock PI, the most-fitting CSV-file (chosen by country) is automatically updated with real-time data from Yahoo! Finance.

CSV Downloader

You can also download CSV's containing stock tickers fairly easily yourself with csv_downloader.py. In this file you will find the code that I used to download the ticker-CSV's from EOD Historical Data into cache/. You can use it as a guide.

Structure

Flow Diagram

Known Issues and Troubleshooting

  1. Issues related to Yahoo! Finance

    • When executing data_into_csv.py I sometimes received the following errors:

      • TypeError: 'NoneType' object is not iterable
      • ConnectionError: HTTPSConnectionPool(host='finance.yahoo.com', port=443)

      Running the script again always fixed it for me.

    • Sparse CSV files

      After sending a large number of requests in a short time period to Yahoo! Finance, one's IP address gets temporarily banned and Yahoo! Finance sends back empty data. This is why the CSV files get more sparse towards the end.

  2. I tried my best to catch errors related to user-inputs. Also the user should always be able to abort the running main.py by typing break. Sometimes this throws weird errors after abortion.

  3. Because the cache/ was pre-downloaded and is not regularly updated, I cannot guarantee for the accuracy of the cached data.

Sample Stock Tickers

  • AAPL for Apple, Inc.
  • MSFT for Microsoft, Inc.
  • BABA for Alibaba Group Holding, Ltd.
  • You can search for any public company on Yahoo! Finance and it's ticker will appear next to it's name.