Gas Emission Level and Stock Market in Energy Sector

Z-score Normalization Formula
Energy sector and number of stocks for each category
Pearson correlation energy sector stocks’ values vs gas emissions
Stock vs Gas level (EMR - HCFC)
N2O, CO2, and CH4 levels from 2001 - 2021 in the United States
Correlation matrix of energy sector stocks’ values vs gas emissions
Uranium normalized stock value with 30 rolling averages
URANIUM Stock Prediction for Energy Sector
Uranium vs CO2
Thermal Coal normalized stock value with 30 rolling averages
Detrend example of Thermal Coal
Oil and Gas Drilling normalized stock value with 30 rolling averages
Normalization levels of the three stocks
Normalization levels of all the gases

Inspiration

We want to see if there's any connection between the environmental effect

How we built it

In this challenge, we are provided with 2 data sources: Knoema’s Data Atlas and Zepl’s US Stock Market for Data Science. Our mission is to find the connection between these two data sources, seeing if there’s a visible impact among environmental factors and stock prices. The challenge required us to work with Snowflake database to derive meaningful insights. We used Python connector library at first to connect our IDE with the database. After that, we downloaded all the necessary files and started analyzing to find the pattern.

Background

We first explored the data with basic and find the unique attributes of the columns from all the datasets. We found out that there’s no linking attribute between stock and environmental factors. Therefore, we decided to check the correlation between gas emissions (primarily N2O, CH4 and CO2) and stocks in the US. To do that, we normalized environmental data and stock data with “Z-score normalization” technique, based on mean and standard deviation values of the data.

After that, we created a function to merge environmental and stock datasets together and plot them in a graph to check their correlation. We formed our hypothesis for the datasets: “There’s a correlation between Energy sector and the gas emissions level in the United States.”

To prove our hypothesis, we explored some factors:

Energy sector and number of stocks for each category
N2O, CO2, and CH4 level from 2001 - 2021 in the United States
Stocks by industry in energy sector Some interesting facts: Uranium Bubble in 2007 Peak in Methane consumption 2009 Global coal benchmarks fall below 2009 crisis levels
Correlation between energy sector stocks’ values vs gas emissions Since there are missing data in the Date and Year attribute from gas emission values, we interpolate the data and used linear regression to detrend and Pearson’s correlation table to get the final score. Finally, we created a matrix to fully recognize the correlation between values.
Visualize to find the industry and gas emission that has the highest correlation
Built machine learning model to forecast the stock market for Uranium

Challenges we ran into

The data resolution for gas emission is yearly while stock is daily. Therefore, many interpolation and data cleaning were implemented
It is hard to choose the right group of stock to study from since stock market is influenced by many factors rather by solely environmental impact.
Grouping stocks within industry require normalization and a lot of trial and errors on data aggregation.

Accomplishments that we're proud of

We were able to download the data from SnowFlakes and figured the focus group to study.
We figured out the correlation and also had time to implement a machine learning model.
We are also find enjoyment in collaboration with teammates

What we learned

The gas emission level has changed dramatically throughout the past decades due to government regulation and social awareness.
There is a slight correlation between CO2 stock and Uranium.