Global Carbon Footprint (1970 - 2015), Carbon Emission
Project GitHub | LinkedIn: Niki Moritz Hao-Wei Matthew Oren
The skills we demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
This project is an interactive and dynamic visualization tool that displays global Carbon emissions data over the years (from 1970 till 2015). It is hosted on a Shiny application that was developed in RStudio.
Please check out the app here & the code behind the project can be viewed here.
Even though the history of climate change science reaches further back, one of the earliest recognizable and significant actions taken by the global community to address climate change from artificial human activity was the United Nations Framework Convention on Climate Change (UNFCCC) established in 1992.
In fact, the Kyoto Protocol which was the result of UNFCC negotiations was not adopted until 1997. Since then, a larger discussion on climate change has taken root in global media and culture and has truly become widespread. Several attempts to reach a global consensus on the courses of action necessary to remedy this increasingly urgent issue has accumulated into the Paris Agreement in 2015.
I was curious in our progress (as the human race) on this existential issue, and quite naturally, the years 1992 and 2015 felt significant as markers in this evaluation. Fortunately, I was able to find a dataset that complied Carbon emissions data on roughly the same number of years before and after 1992.
More specifically, I was interested in seeing chronological relationships between population size and carbon emissions and to see how those trends varied over nations globally.
Looking to bring the numbers to life, the purpose of this project isn't necessarily to provide inferential information through analysis but rather to provide a visual insight into the story the data is trying to tell - the views that are not available through the rows and columns of .csv files.
The data source itself, however, consists of 190 countries and their carbon emissions data through 6 different metrics that are categorized over production-based and consumption-based emissions. Also available in the data set are the population and GDP records.
The data for this project is from the worldmrio.com website. The Eora site describes itself as a "global supply chain database" that "consists of a multi-region input-output table (MRIO) model". The particular data set is structured with 1,522 observations and 47 variables.
The complete information and data are available here.
Project Application (ShinyApp)
The application is designed to be intuitive for the user by communicating insights into wide-ranging and complex data through concise and direct visualizations. The Shiny Dashboard format was used to provide various ways to look at the data set.
The following is a screenshot of the home page of the application:
Beyond the home page, the application consists of useful aggregations of the data, outlined by the page tabs on the left. This is where the project focuses on its story-telling capabilities.
Ultimately, the perspectives extracted from the visualizations are limited only by the user's ability to analyze the information and the external, corroborating knowledge available. The next step within a Data Science approach would be statistical analysis to generate quantifiable inferences and perhaps even further into predictive modeling. But prior to those efforts, this application or tool essentially tries to enable preprocessing or EDA (Exploratory Data Analysis) as an initial step.
Here are some interesting trends I was able to observe from the application:
1. Global Carbon Emissions through Consumption(in GgCO2) in 1995 and 2015:
The global map representation of the data by year reveals trends that can be observed/ reinforced by trade, production, and population rates globally. We see that in 1995, the United States outpaces China in consumption even though the population is lower. However, several years of sustained economic growth has enabled the average Chinese consumer to attain more products and thereby increasing their amount of emissions from consumption. The trend from 1995 is reversed in 20 years by 2015.
2. Global (Aggregate) Carbon Emissions (in GgCO2) from 1970 - 2015: Consumption vs. Production:
A much larger trend that we are able to notice through the project are the differences in global aggregate emission levels from consumption versus emission levels from production per unit of GDP (Gross Domestic Product) from 1970 until 2015.
The steady rise in aggregate, nominal emission levels from consumption can be attributed to growing global population rates as well as an overall increase in access to technology and products. However, if we compare this trend to emission from production as a rate of GDP (if we were to consider GDP as a measure of productivity in global commerce), we see an aggregate downward trend. This is interesting and perhaps attributable to the increasing efficiency of our current technologies (transportation, industrial etc.)
3. Differences in Consumption vs. Production (carbon emission levels) between the United States and China from 1970 until 2015 :
As an additional example of the type of trends that can be observed through this project, the third page allows the user to see consumption vs production emission levels of each country. If we compare the US and China (as the largest two economies and also polluters) we see certain trends become apparent.
A brief inspection of China shows a steady and uniform rise in both consumption and production based emission levels, however, the emission from production is consistently higher. This is clearly indicative of the increasing manufacturing economy developed in China, as well as their growing consumer rates.
Similarly, inspecting the US chart reveals the increasing service economy that has developed in the United States where consumption levels are higher than production emissions. Additionally, we see a decrease in overall rates starting around 2005. This is an encouraging trend, perhaps indicative of a cultural and policy shift.
There is definitely room for improvement and expansion. I would like to apply some refined statistical analysis to the data to glean better insights and inferences. Since the completion of this project, my data science knowledge and experience has definitely improved and I am looking to apply those changes to this project as well!