The 50th Anniversary of the first Earth Day is this week, on April 22. The idea of the original Earth Day was that human health and planetary health are tightly linked: to protect one, we must protect the other. What might our planet be like 50 years from now, on the 100th Anniversary? In this post, I’m putting out my vision of that 2070 future, and some steps we can take now to get there, even in the midst of the COVID-19 crisis.
Category: Data Science
In mid-March I attended a talk organized by the ClimateLink SF Meetup, titled “Rise of the Machines: AI for Energy Management.” The three presenters were Daniel Lim, Co-Founder at Crocus Energy, Pratik Parekh, Sr. Data Scientist and PM at Bidgely, and Andrew Byrnes, Investment Director with Comet Labs. In this post, I’ll summarize the main points of the talks and their connections to data science.
We’ve been exploring a problem in image classification applicable to land use management. This is the fourth and final post in a series that also illustrates the application of the CRISP-DM framework. For the more technical portions of this post, I again give huge kudos to Fast.ai for their technology and courses. Here is the Kaggle notebook1 that goes with this post. Note that fast.ai is a fast-moving library, and some of the functions have changed since my last post; so that code does not fully work anymore!
As I mentioned at the end of the last post, there are many opportunities to improve the model. Some of these are explained in fast.ai lessons 1 and 3, and many more are mentioned in Kaggle forums. I’m not covering those here, but mention a few possibilities briefly.
Welcome to the next post in my ongoing series covering image classification for land use management, a key driver in carbon sequestration. We’re approaching this problem using the CRISP-DM methodology, and have reached step four, modeling. We build up from a simple to a more complex model, and discuss how we will assess the models. We’ll cover their full evaluation in a future post. The code for this post is divided into this Jupyter notebook and this Kaggle kernel, for reasons we discuss.
Baseline Model
In my last two posts I discussed classification of land use and weather in Amazon images. This week, I’m taking a break in that series to cover a recent event I attended at Stanford University. I’m lucky to live near Stanford, which hosts a diverse array of free or inexpensive talks and events. I frequently attend the Energy Seminar, sponsored by the Precourt Institute for Energy. Last week I attended a panel called “Devices and Decisions for Smart Buildings,” moderated by Strategic Energy Alliance Executive Director Richard Sassoon. The three panelists were Anthony Diamond, CTO and Co-founder of Axiom Exergy; Matt Duesterberg, CEO of OhmConnect; and Mark Chung, CEO and Co-founder of Verdigris. I was excited to learn more about this industry and am eager to share some insights with my readers. I’ll cover each company’s offering, use of data science, and future directions.
In my last post I introduced a problem in image classification for land use management. We’re looking at images in the Amazon as provided by Planet. I also outlined the CRISP-DM methodology, covering step one, “business understanding” and step two, “data understanding.” Today I will cover “data preparation;” the goal of this step is to get the data ready for modeling, which is the fourth step in the methodology. I’ll show only code snippets here – the full notebook can be found here.
Data preparation is a bit unusual for this problem, for two reasons. First, we have a multi-label classification problem: each image can be labeled with more than one class. Second, we are dealing with image data, whereas in the majority of data science problems we’d have a variety of numeric and categorical variables. Let’s cover these in turn.
In my Drawdown examples post, I mention the agriculture sector: The emission category of agriculture, forestry, and other land use accounts for 24% of greenhouse gas emissions. In a similar vein, Drawdown’s analysis says the Land Use sector is third in terms of its existing emissions reductions solutions potential; Food, which includes agricultural solutions, has the largest potential. Understanding the current state of the land can help decision makers better manage afforestation, restore and protect forests, and implement agricultural solutions. Data science can help with this understanding, using the tools of image processing and automated classification. In this post I’ll walk through some first steps in applying these tools. I’m assuming you know some Python, but even if you don’t you should be able to follow along.
After reading my first post, you might be wondering, what do artificial intelligence (AI) and data science (DS) have to do with reversing global warming? I’ll start to answer that question in this post with some examples. And of course future posts will include many more examples, and go into more depth.
Sectors
Many emissions analyses divide the world into economic sectors. The EPA publishes data on global emissions by sector based on analysis by the IPCC, and summarized in the figure below. Let’s examine some Drawdown Data Science examples for two large sectors, noting that the pie chart below is based on 2010 emissions.
Welcome to my blog!
The Drawdown Data Science blog is about reversing human created global warming, with a focus on data science and artificial intelligence techniques as significant pieces of the solution. I am motivated to communicate because of my goal to act to solve climate change. I intend to spur others to contribute to solving this global crisis. And yes, it is nothing short of a crisis. There is nothing to debate here, when 97% of climate scientists agree that humans are causing global warming.
Global warming and the associated climate change is happening already — in your state, town, district, or country. It’s not good for anyone — whether it’s hurricanes, drought, fires, extreme heat, extreme storms, flooding, or just a shift in the growing season near you…these things have all been influenced by the rising temperatures. But there are plenty of resources discussing these changes. I direct you to a few in the related resources section of this blog.