Seth Guikema

@SGuikema
55 Followers
131 Following
13 Posts
Professor of Risk Analysis at University of Michigan , Professor II of Risk Analysis
at University of Stavanger. Dad, runner, cyclist, backpacker. Opinions mine. he/him/his. 🇺🇦🏳️‍⚧️
Webpagehttps://guikema.engin.umich.edu
LinkedInhttps://www.linkedin.com/in/seth-guikema-27536b25/
Scholar.Googlehttps://scholar.google.com/citations?user=sYAtnnMAAAAJ&hl=en&oi=ao
Seth GuikemaDec 22, 2022
The storm hasn't even started in Ann Arbor yet, but we just had our first power outage. Outages with wind chills well below zero are likely with the wind that is coming. Be ready!
Seth GuikemaDec 22, 2022
Earlier this year Tessa Swanson, a PhD student I am working with, led a paper outlining the challenges associated with trying to use existing covid data for AI-based prediction of pandemic trajectories. https://doi.org/10.1080/13669877.2022.2077411. The data have only gotten more problematic as testing has been curtailed. Treat any AI-based prediction of covid case as suspect until you really understand the underlying data in detail and trust it. An AI model is not better than the underlying data!
COVID-19 has illuminated the need for clearer AI-based risk management strategies

Machine learning methods offer opportunities improve pandemic response and risk management by supplementing mechanistic modeling approaches to pandemic planning and response based on diverse source...

Taylor & Francis
Seth GuikemaDec 22, 2022
In another 2022 paper with colleagues (https://doi.org/10.1021/acsestengg.2c00008) we developed a utility-theory based approach for ranking areas of a water utility service area to help prioritize lead service line replacement. This paper was a great collaboration across different intellectual areas, and the ranking approach provides a much stronger theoretical basis for ranking areas by vulnerability. And yes, our results show that the approach used for ranking can matter a lot.
Seth GuikemaDec 21, 2022
In another 2022 paper published with my former student Tom Logan and colleagues Roger Flage and Terje Aven (https://www.nature.com/articles/s41893-022-00893-w) we argued that we need to take an integrated view of risk and resilience. A modern interpretation of risk science suggests that risk and resilience are not separate, and we should be taking an integrated view rather than setting up artificial walls between the concepts of risk and resilience.
Risk science offers an integrated approach to resilience - Nature Sustainability

In the face of growing calls to restrict risk analysis to narrow and specific events, this Perspectives argues instead for fully integrated frameworks that bring risk analysis into all aspects of resilience studies.

Nature
Seth GuikemaDec 21, 2022
While in COVID isolation, I am reflecting on papers I co-authored with students in 2022. The first is https://doi.org/10.1111/risa.13955, a part of Chengwei Zhai's PhD dissertation. This paper used an agent-based model to assess the potential for household learning to change the trajectory of community vulnerability. Prior beliefs about storm risk can play a large role in post-hurricane mitigation decisions, and individual mitigation can change overall community building stock vulnerability substantially.
Seth GuikemaDec 20, 2022

Are you an academic? Please strongly consider saying yes when asked to review a paper. I am an editor in two journals, and it is getting increasingly challenging to find reviewers. I am now routinely asking 20-30 potential reviewers to find 2.

A rule of thumb I was taught in graduate school was that for each paper you submit to a journal, you should be willing to say yes to 3 reviews. I still think this is a good rule of thumb! Strong peer review is essential.

Seth GuikemaDec 18, 2022

Trying get help from Xfinity at Christmas time:

800-number: Please use the online chat bot because of high call volume.

chat-bot: Please call 1-800-xfinity due to high customer volume.

They've perfected the infinite loop.

Seth GuikemaDec 17, 2022
Emily FarrisDec 17, 2022
Don’t compare yourself to others in academia in productivity. You don’t know their teaching load, their research support, their institutional resources, etc. You cannot win in this game.
Seth GuikemaDec 16, 2022
Hmm. This is deeply problematic.
https://www.vice.com/en/article/dy79bx/twitters-new-head-of-trust-and-safety-offers-to-partner-with-controversial-anti-trafficking-group
Twitter’s New Head of Trust and Safety Offers to Partner with Controversial Anti-Trafficking Group

Operation Underground Railroad has a long track record of misrepresenting its work; now, it could be partnering with Elon Musk’s Twitter. 

Show threadSeth GuikemaDec 15, 2022
As we wrote earlier in the pandemic, Covid data is deeply problematic (https://doi.org/10.1080/13669877.2022.2077411). This isn't due to some conspiracy or malfeasance. It's a combination of how testing and the pandemic have evolved together with the steadfast refusal of agencies to institute a statistically rigorous community sampling program to try to estimate prevalence. Closest we have is sewer monitoring, and in many places that is suggesting a much larger increase in cases than the problematic testing data is.