The story behind WeatherAlarm
The idea for WeatherAlarm started long before the app itself existed.
In 2008 I graduated from the Faculty of Computer and Information Science. Soon after, I built my own business management software, which is still in use today by roughly 50 companies.
Alongside that business software, I had been carrying a slightly unusual idea for a long time: to build a program, an algorithm really, that could warn you about approaching precipitation before it reached your location.
Back then the idea was limited far more by technology than by willingness. Smartphones with GPS, a permanent internet connection and enough computing power simply did not exist in a form that made such a system practical.
With the arrival of smartphones, the idea became feasible.
The first sketches
In 2010 I started drafting the first outlines of the system and thinking about how to work out, from radar images, where the precipitation was heading.
The core idea was simple:
I do not only want to know whether it is raining right now, but whether the rain is coming towards me.
Showing a radar image is not enough for that. The system has to analyse a sequence of radar frames, identify precipitation cells, work out their direction and speed, and then calculate where they will be after a given time.
Development begins
In 2011 I began building the first version of the Android app.
Most of the work was not on the phone. The phone was mainly the user interface; the demanding part ran on the server.
Step by step, a system took shape that tried to derive from radar data:
- the direction a precipitation cell is moving,
- how fast it is moving,
- how intense it is,
- its expected path,
- the approximate time it will reach the user.
That meant considerably more work than building an ordinary weather app. It was not a matter of displaying data prepared by some other weather provider. I had to develop a large part of the data processing and movement forecasting myself.
The first public release
In 2012 the first version of WeatherAlarm was published on Android.
That first version was of course far simpler than today's, but it already carried the core idea that has stayed the same ever since:
warn the user when precipitation is genuinely approaching their location.
The release date of the first Android version can still be checked in independent app archives, which list 15 February 2012 as the date of first publication.
Growth and the first users
Over the following years I gradually improved the system, added new regions and refined the algorithms.
The app slowly gained users and ratings. By 2016 it was widespread enough that Google Play archive data recorded more than 10,000 installs.
The most important part of that period was not only the growth in user numbers. I kept testing the system against real weather events and adjusting it according to the results.
Development gets more serious
In 2017 I went back to developing the app more intensively and presented it on Slovenian and international technology and weather forums.
On Android Central I described how the system worked: the app used official radar imagery, while the calculations ran on my own server. The algorithm tried to predict the direction, speed and intensity of precipitation roughly two hours ahead.
By then the app was available for Slovenia, Croatia and the United States.
The users themselves became an important part of development. Their comments, suggestions and reports from the field helped improve the app further.
Expansion and years of development
Over the following years WeatherAlarm gradually expanded to other regions and gained an iOS version.
Development ran alongside the rest of my work, so WeatherAlarm was never a classic large development team project. I built most of the system myself, from the app to the server side and the algorithms.
By 2023 the app was still actively published and had around 1,000 ratings on Google Play; the last release of the older generation came out in July 2023.
A pause
After more than a decade of development came a period when I did not give the project as much time.
The app stayed published, but development did not continue at the same intensity.
The idea, however, stayed the same.
A fresh start
In 2026 I decided to develop WeatherAlarm actively again.
This time not merely as an update of the old app, but as a considerably larger project.
New features joined the core radar algorithm, among them a more detailed view of how precipitation moves, an estimated arrival time, hail warnings and live community reports from users.
Today a user can report hail, heavy rain, wind, lightning or flooding, and other users nearby can see that event on the map.
User photographs, a forecast history and weather prediction games were added as well.
Today's version is therefore the result of more than 15 years of work on the idea, from the first sketches in 2010 to a system that now combines radar data, my own algorithms and data from users.
What it looks like now
In August 2026 the home screen shows the precipitation level at your location, the distance to the nearest cell and a graph of the next 60 minutes. The map adds the forecast paths of the cells, their speed, lightning and reports from users.
From an idea to a system
Looking back, what strikes me most is that the core idea has not really changed in all these years.
In 2010 I wanted to answer a very simple question:
"Is rain or a storm coming towards me, and when will it arrive?"
Today WeatherAlarm tries to answer that question with far more data, more capable algorithms and the help of a community of users.
These days artificial intelligence gets mentioned at every turn. There is no AI detection of precipitation in this app. It is an algorithm I thought up more than 15 years ago and have been refining on real weather events ever since.
So for me the project did not begin in 2026.
In 2026 it simply got a fresh start.