Machine Learning and Epidemiology with Elaine Nsoesie - #396

Surveillance Methods

Elaine Nsoesie discusses the use of machine learning in health surveillance, emphasizing real-time monitoring techniques. By collaborating with local health departments, she developed platforms to analyze social media posts for signs of foodborne illnesses, enabling targeted inspections and outbreak investigations 1. This proactive approach helps identify potential health threats before they escalate. Elaine also highlights the importance of contextualized approaches in Africa, where search behaviors differ from Western countries, necessitating tailored interventions 2.

When public health is working, we don't really think about it. It's when we have major outbreaks and we see things happen in our communities that we start thinking about it.

--- Elaine Nsoesie

Additionally, she explores the potential of using Amazon reviews to detect unsafe food products, aiming to prevent consumption-related illnesses 3.

   

Satellite Imagery

Elaine explains the innovative use of satellite imagery in health analytics, particularly for tracking diseases and predicting health outcomes. By analyzing hospital parking lot usage through satellite images, her team could detect trends in influenza-like illnesses, offering a novel method for disease surveillance 4. This approach provides valuable insights, especially in regions lacking traditional data sources.

We were able to show that you could actually capture trends, influenza-like illnesses during the flu season by just looking at how people were using hospital parking lots.

--- Elaine Nsoesie

Elaine's work also extends to assessing neighborhood health indicators, using satellite data to predict obesity prevalence based on environmental factors 5.

   

Predictive Analytics

Elaine employs advanced modeling techniques to understand health phenomena, focusing on obesity and infectious disease spread. Using random forest regression and neural networks, she identifies predictive factors of obesity, demonstrating the influence of environmental elements on health 6. Her research underscores the importance of comprehensive data analysis in public health.

We use transfer learning and convolutional neural networks to process the images.

--- Elaine Nsoesie

Elaine also examines COVID-19 disparities, highlighting how socioeconomic factors affect disease spread and response 7.