A couple of weeks ago, MSEM and GsAL hosted a panel discussion on the role of data analytics and machine learning in environmental management. We invited Jeff Dlott, Josh Strauss, and Andrew Notohamiprodjo to provide their insights from their respective spaces within the private sector. To open the event, we asked each panelist to share a project in which they utilized data-driven technologies to innovate the way they approached answering environmental questions.
Jeff Dlott, Chief Operating Officer at Landscan.ai, shared with us an overview of Landscan’s portfolio of projects that focus on helping farmers manage agricultural lands to optimize their yield, while maintaining plant and soil health. Jeff brought with him one of the physical probes that Landscan has developed to measure in-situ soil health variables, like moisture, nutrients and microbial ecology. Landscan utilizes machine learning to develop models that can predict plant health and crop yield by combining rich, remotely-sensed datasets collected above ground and in-soil subsurface data, collected and analyzed on the spot. Jeff explained how their state-of-the art technology for in-situ measurements surpasses the accuracy and efficiency of standard lab-based practices in assessing soil health, and increases the accuracy of their predictive models.
Josh is the President of Environmental Products at Anew, a company focused on accelerating the transition to a low-carbon economy through innovative environmental solutions and sustainability projects. Anew’s portfolio includes the calculation of carbon credits that can be achieved through forest management strategies. Josh explained how, by employing machine learning techniques, Anew scientists use dynamic baselining to calculate carbon credits enabled by machine learning-based remote sensing models. These models allow the calculation of land management benefits by comparing to actual outcomes from comparable properties from a rich database and a range of data layers, resulting in a much more nuanced and accurate prediction of the benefits of proposed forest management practices.
Andrew Notohamiprodjo, MSEM ‘19 alum, now the Head of Data Science at Delos Insurance Solutions, discussed the pivotal role that advanced fire-risk modeling has played in improving the ability of Delos to provide insurance to homeowners. By leveraging satellite imagery, weather data, machine learning and state-of-the-art science on how fires behave in the new era of climate change, Andrew and the Delos team can assess fire risk at a much higher resolution, and at a greater accuracy than the models that have historically been used by wildfire insurance providers. Great emphasis was placed on the need to be utilizing the best available science in the creation of new tools in the industry, and to ensure that they are being developed for the purposes of the questions that are being asked. If you want to learn more about Andrew’s work, check out his recent interview with Insurance Business.
After getting to hear about the work the panelists are involved in, and the critical role that machine learning plays in that work, we opened the floor to questions from the audience. One of the most provoking questions that was asked pertained to the role that machine learning plays in regards to environmental justice: Can these models be trusted to provide accurate results, and to not perpetuate environmental injustices? In their responses, the panelists spoke to how machine learning tools, in the scope of their projects, are not used as black boxes. Rather, they have domain experts from all relevant disciplines work together with computer and data scientists to provide the most accurate predictions, while keeping important, multi-faceted objectives like environmental justice in the conversation.
The event closed with the panelists providing their insight on how students can be best prepared to enter the profession. Especially in this new, highly technical, highly interdisciplinary and, heavily data-driven era of environmental management. Here is what they had to say:
- Be a Domain Expert and Collect Multiple Skills: Now, more than ever, it is important to enter the profession with an expertise and deep knowledge of a selected field. However, they also noted that you should be ready and willing to work within areas that are outside of your expertise.
- Communication and Collaboration Skills are Important: The field of environmental management is the most interdisciplinary it has ever been. In every sector, teams of multiple experts work together to solve difficult problems, so strong communication skills are more important than ever. They encouraged students to get as much practice as they can while they are completing their studies!
- Asking Good Questions is Better Than Attempting to Know All the Answers: All panelists highlighted that the environmental management field is looking for early career seekers who are motivated to work and learn, and who are genuinely passionate about the mission of the companies they work for. They also noted that companies do not expect new hires to know everything, and that they actually want folks who will ask good questions and employ critical thinking as they learn their roles.
To provide insight from a student perspective, ENVS ‘23 alum Kevin Galvin was invited to join the panelists for closing statements. Galvin recently completed an internship with the Environmental Defense Fund in San Francisco, in which he worked to develop methods for capturing methane releases from the oil and gas industry using machine learning and remote sensing. He spoke to how his educational path, which included a GIS certificate from USF’s Geospatial Analysis Lab (GsAL), was pivotal for his current research.
Thank you to all of our panelists for their insightful comments, and to our audience for their engagement and curiosity on this vital topic! We look forward to hosting events similar to this during the Spring 2025 semester, so be on the lookout for updates on programmatic events.
