Copyright © 2018 DataScience.US All Rights Reserved.
Using Big Data and Machine Learning to Fight Global Warming
Climate change is one of our world’s biggest and most complex problems.
Given the scope and sheer importance of the problem it makes sense to be using every tool we have to combat it. The tools available to us include big data and machine learning, and innovative applications for these technologies can help us turn the tide against climate change. Machine learning and big data can assist in fighting climate change by allowing scientists to do things like create more accurate climate models, predict forest fires and other extreme weather events, and assess the biggest sources of greenhouse gases.
The first step in fighting climate change is to get an accurate picture of how our world is changing. There are many different projects which use big data to create more accurate climate models. For instance, NASA continuously collects data on the conditions of land surfaces around the world. This info is aggregated at Landsat, and it has been a critical tool for those who research the effects of climate change on the environment.
NASA has stated that the data from Landsat is critical for determining how human activity is impacting the climate and the earth’s surface. This constantly updated data set allows scientists to create predictive models that can determine what the future of the globe might look like, areas particularly vulnerable to climate change, as well as the best methods to combat climate change. The Environmental Protection Agency has also made use of big data to assess what the largest sources of emissions are and how these emissions impact certain areas of the world.
This form of vulnerability research is critical for informing our decisions about climate change, and as a field of research has expanded dramatically over the last few years, employing new and innovative techniques to assess risk zones for climate change related disasters. Anonymized records from cell-phones such as call details and geospatial data can be used to create a dynamic population map that could track a population’s risk for natural disasters, including variables such as population, economic status, sensitive infrastructure, etc.
Natural language processing can also be used to analyze massive volumes of text data, for example discussions about climate change on social media, to learn about perspectives on climate change. Using natural language processing could reveal which impacts of climate change are the most and least frequently discussed, along with locations frequently categorized as being the most vulnerable.
Calculating Better Solutions to Climate Change
Machine learning can be employed to help fight air pollution and emissions of greenhouse gases. By making predictions about the effects of certain proposed solutions, researchers and policy makers can be better informed about the tactics that would work to reduce air pollution and tactics that wouldn’t work.
IBM Research China is working on a project called Green Horizon, which tracks major sources of pollution and can generate “What-If” scenario analyses. The system is intended to be able to recommend tactics that can effectively reduce air pollution, such as temporary restrictions on the number of drivers who can be on the road. IBM used data from the Beijing Environmental Protection Bureau to construct its models, and the models generated by their machine learning algorithms were 30% more accurate than the models generated by traditional means of calculation.
Machine learning can also be used to make more efficient use of our energy, necessitating less greenhouse gas emissions from power plants. Last year, Google used machine learning algorithms to cut the amount of energy used at their data centers by a full 15%. Data from DeepMind was used to calculate more efficient ways to cool and pressurize server rooms, leading to the drop in energy consumption.
This year saw the UN’s Global Pulse Initiative host their Big Ideas Competition 2017, which focuses on the use of big data to create solutions to combat climate change in the Asia Pacific region. The competition wants participants to create data-driven solutions that will strengthen the resilience of the region against climate-related hazards.
Educating the Public
Big data and machine learning even have roles to play in public engagement and education. An informed and engaged public is critical to combating climate change, and the data collected by Global Pulse, the EPA, and other organizations can be used to reflect public sentiment about climate change. The trending topics related to climate change and the opinion on those topics can be pulled from social media data sets. Having an idea of what topics the public is interested in, and what the opinions on those topics are can make it easier for climate scientists and activists to target their education and outreach attempts. If you want to educate people and change how they think, you must first know what they are thinking.
Ultimately, the key to solving any problem, regardless of its scale lies in being informed about the problem. Big data can inform researchers about global climate change, and machine learning can be used to help determine the most effective ways of combating it.