AvvySavvy

AvvySavvy Empowers safe winter adventures with data-driven avalanche intelligence.

Our Mission

AvvySavvy empowers everyone to enjoy winter sports with confidence by providing:

• Simple, understandable avalanche forecasts
• Real-time risk assessment based on location and conditions
• Educational resources to promote avalanche awareness

Problem

The problem being addressed is the significant threat that avalanches pose to individuals exploring mountainous terrain, resulting in tragic outcomes due to navigation errors, unexpected hazards, and a lack of situational awareness about mountain conditions. Avalanches lead to an average of 30 fatalities annually, with causes ranging from navigation errors to a limited understanding of terrain and hazards. Specifically, navigational errors contribute to the risk, as individuals may find themselves in avalanche-prone areas without adequate knowledge of the terrain, accurate route planning, or awareness of avalanche-prone zones.

Motivation

Mitigating the risks associated with avalanches is compelling and impactful due to the potential to save lives and enhance the safety of individuals engaging in snowsports. The proposed solution, a AvvySavvy App, aims to empower users with real-time location and direction information, precise terrain knowledge, and hyper-local avalanche alerts. The impact of solving this problem extends beyond personal safety; it contributes to the enjoyment and exploration of snow-covered landscapes with peace of mind.

Target Users

Overall, AvvySavvy serves as a comprehensive solution that benefits snowsports enthusiasts, resorts, and residents alike, fostering a safer and more informed winter sports community.

• Snowsports Enthusiasts: This group includes backcountry skiers, snowboarders, and snowmobilers who are passionate about winter sports. AvvySavvy empowers them to make informed decisions about avalanche risk before embarking on their outdoor adventures. By providing access to accurate and user-friendly avalanche forecasts, AvvySavvy ensures that snowsports enthusiasts can enjoy their activities safely.
• Snow Resorts: Resorts play a vital role in the winter sports industry, catering to guests seeking adventure and relaxation on the slopes. AvvySavvy offers resorts a valuable tool to enhance guest safety and optimize ski patrol operations. By integrating AvvySavvy into their safety protocols, resorts can provide guests with real-time information about avalanche conditions, ultimately ensuring a safe and enjoyable experience for all visitors.
• Residents in Snow Areas: Living in avalanche-prone regions comes with unique challenges and risks. AvvySavvy provides residents with valuable insights into current avalanche conditions, allowing them to stay informed and make educated decisions about travel, outdoor activities, and emergency preparedness. By leveraging AvvySavvy, residents can better navigate the potential dangers of living in snow-covered areas, enhancing overall safety and resilience within their communities.

Data Source & Data Science Approach

The AvvySavvy application draws upon a comprehensive array of data sources and employs sophisticated data science techniques to generate accurate avalanche risk predictions. Firstly, it utilizes the Daily Backcountry Avalanche Forecasts and Daily Observations provided by the Sierra Avalanche Center, which offer crucial insights into local snowpack stability, recent avalanche activity, and other pertinent observations from field experts. Additionally, the application integrates data from the National Weather Service API and the Weather Data API by Synoptic Data to access meteorological information such as recent snowfall amounts, wind speed and direction, temperature fluctuations, and precipitation forecasts.

To process this wealth of data and generate predictions, AvvySavvy leverages advanced machine learning models, specifically employing time-series analysis and Feed-Forward neural networks. These models are trained on historical avalanche data alongside corresponding meteorological and observational data, allowing them to identify patterns and correlations between various factors and avalanche risk levels. By analyzing factors such as recent snowfall accumulation, wind loading, temperature gradients within the snowpack, and terrain characteristics, the machine learning models can effectively predict the likelihood and severity of avalanches for specific locations.

The expertise of avalanche forecasters and researchers is crucial in interpreting the data and identifying the most relevant factors contributing to avalanche risk. These experts contribute to the development and refinement of the machine learning models by providing insights into the complex interplay between weather conditions, snowpack characteristics, and terrain features. Through this collaborative approach, AvvySavvy is able to provide users with detailed and accurate avalanche risk assessments, empowering them to make informed decisions and prioritize safety while enjoying snowsports in backcountry environments.

Model

First, users begin their journey by visiting the 'Plan My Trip' page, which corresponds to model 1. Here, users input latitude-longitude details and select their desired date of arrival. The platform then analyzes this information to determine a date that is safe for snow sports activities.

Once users have identified a safe date, they arrive at their chosen spot and proceed to check the avalanche risk for the specific elevation they're interested in. This is facilitated by model 2, which provides detailed risk assessments based on elevation data.

After selecting the desired elevation, users can then check the avalanche risk at the exact spot they're planning to visit. This is made possible by model 3, which offers real-time risk assessments tailored to the user's specific location.

Evaluation

In the AvvySavvy project, our primary objective is to improve safety by accurately predicting avalanche probabilities. We place a higher emphasis on recall rather than precision. Recall measures our model's effectiveness in correctly identifying avalanches when they happen, thus reducing the occurrence of false negatives. False negatives present a serious safety concern as they occur when the model fails to predict an avalanche despite its presence. By prioritizing recall, we strive to minimize false negatives, thereby enhancing safety by ensuring our model reliably detects avalanches.

Impact

The opportunity lies in the size of the global snowsports market, estimated to reach USD 76.8 billion by 2027, with a growing trend of users seeking enhanced safety and technology-driven solutions. The increasing adoption of innovative technologies, such as machine learning (ML) for avalanche prediction and mitigation, further emphasizes the potential impact and demand for a comprehensive solution like the AvvySavvy App.

Acknowledgments

Our team is grateful to our capstone professors Joyce Shen and Kira Wetzel. Their support, advice, and encouragement was instrumental in the completion of our product.