The latest articles on DEV Community by Evgeny Padezhnov (@evgeny_padezhnov). https://dev.to/evgeny_padezhnov https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3879975%2F276e7f6b-3959-437e-996f-afb97a5f9776.png https://dev.to/evgeny_padezhnov en Evgeny Padezhnov Wed, 15 Apr 2026 07:46:11 +0000 https://dev.to/evgeny_padezhnov/how-i-built-an-airspace-risk-monitoring-system-with-ml-and-public-data-3l77 https://dev.to/evgeny_padezhnov/how-i-built-an-airspace-risk-monitoring-system-with-ml-and-public-data-3l77 <p>When 12 Flight Information Regions closed simultaneously in February 2026, <br> thousands of flights were rerouted within hours. I started wondering — <br> what if this data could be processed systematically?</p> <h2> What FlySafe Does </h2> <p>FlySafe is an API that processes public aviation data and outputs <br> a simple number: a 0-100 risk index per flight region.</p> <p>Send a route → get an index. That's it.</p> <p>Updated every 5 minutes. Three time horizons: 72h, 7d, 30d.<br> Backtested on historical airspace closure events.</p> <h2> How It Works </h2> <p>The model identifies patterns in public data that historically <br> correlate with airspace disruptions. </p> <p>It doesn't predict specific events. It doesn't replace NOTAMs. <br> It provides a numerical index as one additional data point.</p> <h2> What I Learned Building This </h2> <ol> <li> <strong>Aviation data is messy.</strong> NOTAMs have no standard format across countries</li> <li> <strong>Conflict data is surprisingly open.</strong> Academic databases provide research-grade datasets for free</li> <li> <strong>The hard part isn't ML — it's data normalization.</strong> Getting multiple sources into a unified pipeline took longer than training the model</li> <li> <strong>B2B is lonely.</strong> No viral loop, no app store, just cold emails and API docs</li> </ol> <h2> Who It's For </h2> <p>B2B product — airlines, flight aggregators, insurance, <br> corporate travel platforms. Not a consumer app.</p> <p><strong>Demo (simulated data):</strong> <a href="https://flysafe.zone" rel="noopener noreferrer">flysafe.zone</a></p> <p>Launched on Product Hunt today — feedback welcome.</p> <p><em>All data derived from publicly available sources. <br> No classified or insider information. No safety advice.</em></p> machinelearning api aviation