I like where your headspace is with your product but I do have some quesitons/concerns.
1) What is your data retention and privacy policies..what's to prevent some malicious actor from using your lat/lon conditions to know if you're home or where you are going. Privacy seems to be the name of the game and with that also comes high-security.
2) What kind of data refresh cycles are you looking at...are the data sources you are using pig-backing off open source APIs or using proprietary data aggregation with some algorithms on top? I find that many open source data repos have poor data quality and very long updates with current data. Also, do you guarantee accurate predictions for an hour/day/week and within that kind of fidelity (grid, county, state, country?)
Those are excellent questions carboncycle, and aspects of our service we have taken great care considering and implementing.
Privacy
We appreciate your sensitivity and are committed to respecting and protecting your privacy. Our service monitors and retains ONLY the most recent location provided by your satellite communication provider (Garmin, Zoleo, etc.) and uses that information location to check against our database of active alerts. We discard that location once our servers obtain an updated location, and the process continues, always only retaining your most current location.
We also retain your location information associated with each alert we send to your device for diagnostic and liability reasons.
State-of-the-art, industry-leading cloud computing providers host our service and the data we collect, and we meet or exceed recommended or required data security practices.
You can also review our privacy policy here -
Privacy Policy | Adiona Alert - https://adionaalert.com/privacy
Safety Alert Sources and Their Reliability
Now, to the more complex and technical question about how and where we source the alerts we monitor and deliver to your device when you are inside an affected area.
We do NOT use "open" or "public" sources, such as media or information aggregators - we go to the source of the alerts ONLY and use vetted and secure data feeds that several government agencies operate.
The bulk of the alerts are weather watches and warnings issued by the U.S. National Weather Service (NWS) and Environment and Climate Change Canada, the Canadian counterpart to the NWS. These alerts range from flood and winter weather watches to severe thunderstorms or tornado warnings. In other regions, these could be flash flooding, wind warnings, frost warnings, etc.
Both agencies are also sources for tsunami warnings. In many regions of the U.S., the NWS also issues avalanche warnings when the conditions warrant it.
Given that our focus is on providing these critical alerts in remote or wilderness areas, away from large population centres, we know that the public weather agency takes great care to offer these safety alerts regardless of the commercial interests that might be a challenge for commercial alert providers, who primarily get their severe weather information from the public forecasters anyways.
In addition, we monitor other safety alerts issued by State, Provincial, Tribal, regional or local public safety agencies that address a comprehensive range of hazards. Those alerts include drinking water advisories, active shooters, chemical releases, dam failures, missing persons, dangerous animals - you name it. The key is that these alerts are issued when the local agency, which has the local knowledge and jurisdiction, believes the public would benefit from being notified of a significant risk in their community.
We source the alerts from several feeds managed by various government agencies, including FEMA in the U.S. and Public Safety Canada, and other feeds like the NWS weather alert feeds.
In all cases, the information and alerting agency is curated and vetted, and the feed is secured to ensure you get trusted and reliable information.
To give you a sense of how comprehensive the alerts issued by public agencies can be, a few weeks ago, we recorded the highest number of active alerts at one time since we launched our Early Access program. At one point on December 18th, we were monitoring nearly 900 active safety alerts, ranging from Blizzard Warnings in Alaska and some parts of Canada to Small Craft Advisories on the coast of southern Florida and Hawaii, not to mention the numerous Rainfall and Flood warnings on the west coast of Canada and the U.S.
One of the objectives of our Early Access program is to learn from our users what information is valuable to them and will help them stay safe in the backcountry.
We are continually reviewing the alerts we issue and what alerts that aren't currently available that would benefit our users. Your feedback will also inform our future development plans for other safety information we want to offer but still need to integrate.
Location Accuracy and Alert Timeliness
The location accuracy of the alerts you receive is determined by the accuracy of the location reported by your Garmin, Zoleo or other device, which is dependent on the quality of the GPS signal but is generally accurate to between 10-100’ and the update frequency you’ve set for your device. Suppose you travel quickly, such as in a vehicle, and have long location update intervals. In that case, your location accuracy will be much lower than if you are on a human-powered adventure with a frequent update interval.
Regardless, we will use your location and compare it to the affected area determined by the alerting agency. If you are in the affected area when the alert is issued, we will immediately send you the alert. At the same time, if your travels take you to an area where an alert was previously issued, we will send you the alert as soon as your location is updated by your device and our service determines you are inside the area subject of the alert.
As a result, the two main factors affecting the timeliness of the alerts you receive are the frequency of the location updates you set for your device and the delay in the satellite communication system. There are also some data processing delays in receiving the alerts from the alerting agency, our data processing logic, and the location matching process, but they are essentially immaterial to delaying the notification process.
In practice, I’ve seen alerts received by our test units as quickly as two or three minutes after the NWS issued them to other circumstances where the satellite link was more challenging because of tree cover or the unit sitting on the dash of a car when the message was delayed fifteen or twenty minutes.
Lastly, the alert type is an essential element of timeliness - last week, during the monster waves event in Southern California, the warnings were issued by the NWS hours and received by one of our early users well before the hazardous conditions occurred, providing plenty of time to take protective actions. Other alerts are issued for much more dynamic conditions, such as severe thunderstorms that lead to an outbreak of tornado activity. In such a case, a severe thunderstorm watch would likely be issued hours earlier to the severe weather outbreak, followed by appropriate warnings issued from 15-20 minutes to an hour or more before the “arrival” of the storm, then followed by upgraded warnings, including an observed tornado warning several minutes or more before a storm arrives in the affected area. Such a storm is more challenging for the alerting agency, us, and our users, but this highlights the importance of severe weather watches issued ahead of a period of severe weather.
I’ve attached an excellent image from the NWS that might help you understand how the affected geographic area of a severe weather warning might look to help you get a sense of what we are using.
I wish I could provide an “easy” answer to the technology and science behind severe weather and other safety alerts. Still, it’s a complex topic where the effort is about risk reduction.
I hope this helps you understand the service we are rolling out. Let me know if you have any questions.
J.S.