Unifying Environmental, Social, and Governance (ESG) Metrics with Financial Analysis

DANL 210 Project

Author
Affiliation

Byeong-Hak Choe

SUNY Geneseo

Published

January 22, 2025


Project

  • Publish the webpage of your data analysis project on your website, hosted on GitHub.
    • Your data analysis should focus on the agenda, “Unifying Environmental, Social, and Governance (ESG) Metrics with Financial Analysis”.
    • Use either a Jupyter Notebook or a Quarto document to present your work.
    • The due for the project is May 16, 2025, Friday, 11:59 P.M.


Project Data

  • Below is the esg_proj DataFrame, which provides a list of companies and associated information
import pandas as pd
esg_proj = pd.read_csv("https://bcdanl.github.io/data/stock_esg_list.csv")

Variable Description

  • Symbol: a company’s ticker;
  • Company Name: a company name;
  • Sector: a sector a company belongs to;
  • Industry: an industry a company belongs to;
  • Country: a country a company belongs to;
  • Market Cap: a company’s market capitalization as of December 20, 2024 (Source: Nasdaq’s Stock Screener).
    • A company’s market capitalization is the value of the company that is traded on the stock market, calculated by multiplying the total number of shares by the present share price.



Project Tasks

Data Collection

yfinance

  • Utilize the yfinance library, an unofficial API for accessing Yahoo! Finance data.
    • For each company listed in the esg_proj DataFrame, you should retrieve the following:
      • Daily historical stock data spanning from January 1, 2024, to March 31, 2025.
      • Quarterly income statements covering the period from March 31, 2024, to March 31, 2025 (encompassing five quarters).
      • Quarterly balance sheets for the same period as the income statements.

selenium

  • Employ the Python selenium library to gather ESG Risk Ratings, along with the Controversy Level from the Sustainability Section of each company’s webpage on Yahoo! Finance, such as:
  • Ensure robust error and exception handling during data collection from the Sustainability Section on Yahoo! Finance:
    1. To avoid being blocked by Yahoo! Finance for automated browsing activities, please manage the Chrome driver by launching and quitting it after processing every 20 companies’ Sustainability Sections.
    2. Use time.sleep() for each visit of the Sustainability Section webpage.
    3. Some companies may lack available data on Environmental Risk Score, Social Risk Score, Governance Risk Score, and/or Controversy Level.
    4. Yahoo! Finance intermittently switches between its new and classic website layouts, affecting the HTML DOM structure of the Sustainability Section.
    • You can check the new and classic layouts of Yahoo! Finance by selecting the corresponding options from the navigation bar on their website.


Data Analysis

  • Below are the key components in the data analysis webpage.

    1. Title: A clear and concise title that gives an idea of the project topics.

    2. Introduction:

      • Background: Provide context for the research questions, explaining why they are significant, relevant, or interesting.
      • Statement of the Problem: Clearly articulate the specific problem or issue the project will address.

    3. Data Collection: Use a Python script (*.py) to write the code and the comment on how to retrieve financial, accounting, and ESG data using Python yfinance and selenium.

      • Do NOT provide your code for data collection in your webpage. You should submit your Python script for data collection to Brightspace.

    4. Descriptive Statistics

    • Provide both grouped and un-grouped descriptive statistics and distribution plots for the ESG data and the finance/accounting data
    • Provide correlation heat maps using corr() and seaborn.heatmap(). Below provides the Python code for creating a correlation heatmap.
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Sample DataFrame with varied correlations
data = {
    'Revenue': [100, 200, 300, 400, 500],  
    'Profit': [20, 40, 60, 80, 100],       
    'n_Employee': [50, 45, 40, 35, 30], 
    'n_Customer': [10, 11, 12, 13, 14]  
}

# Create a DataFrame from the dictionary
df = pd.DataFrame(data)

# Calculate the correlation matrix of the DataFrame
corr = df.corr()

# Set up the matplotlib figure size
plt.figure(figsize=(8, 6))

# Generate a heatmap in seaborn:
# - 'corr' is the correlation matrix
# - 'annot=True' enables annotations inside the squares with the correlation values
# - 'cmap="coolwarm"' assigns a color map from cool to warm (blue to red)
# - 'fmt=".2f"' formats the annotations to two decimal places
# - 'linewidths=.5' adds lines between each cell
sns.heatmap(corr, annot=True, cmap='coolwarm', fmt=".2f", linewidths=.5)

# Title of the heatmap
plt.title('Correlation Heatmap with Varied Correlations')

# Display the heatmap
plt.show()

  1. Exploratory Data Analysis:
    • List the questions you aim to answer.
    • Address the questions through data visualization with seaborn (or lets-plot) and pandas methods and attributes.
  2. Significance of the Project:
    • Explain its implications for real-world applications, business strategies, or public policy.
  3. References
    • List all sources cited in the project.
    • Leave a web address of the reference if that is from the web.
    • Indicate if the code and the write-up are guided by generative AI, such as ChatGPT. There will be no penalties on using any generative AI.
    • Clearly state if the code and the write-up result from collaboration with colleagues. There will be no penalties for collaboration, provided that the shared portions are clearly indicated.


Rubric for the Project

  • Here is the link to the PDF file of the rubric for the project web-page.

Rubric for Quality of Data Collection

Evaluation Description Criteria
1 (Very Deficient) - Very poorly implemented
- Data is unreliable.		 - Ineffective use of yfinance, resulting in incomplete or inaccurate financial data.
- Poor web scraping practices with selenium, leading to unreliable or incorrect data from Yahoo Finance.
- Inadequate use of pandas, resulting in poorly structured DataFrames.
2 (Somewhat Deficient) - Somewhat effective implementation
- Data has minor reliability issues.		 - Basic use of yfinance with minor inaccuracies in data retrieval.
- Basic web scraping with selenium that sometimes fails to capture all relevant data accurately.
- Basic use of pandas, but with occasional issues in data structuring.
3 (Acceptable) - Competently implemented
- Data is mostly reliable.		 - Competent use of yfinance to retrieve most financial data accurately.
- Effective web scraping with selenium, capturing most required data from Yahoo Finance.
- Adequate use of pandas to structure data in a mostly logical format.
4 (Very Good) - Well-implemented and organized
- Data is reliable.		 - Advanced use of yfinance to reliably and accurately fetch financial data.
- Thorough web scraping with selenium that consistently captures accurate and complete data from Yahoo Finance.
- Skillful use of pandas for clear and logical data structuring.
5 (Outstanding) - Exceptionally implemented
- Data is highly reliable.		 - Expert use of yfinance to obtain comprehensive and precise financial data.
- Expert web scraping with selenium, capturing detailed and accurate data from Yahoo Finance without fail.
- Expert use of pandas to create exceptionally well-organized DataFrames that facilitate easy analysis.
Back to top