Context
Most commercial aircraft these days send out a data packet every few seconds across the airwaves, and this data contains various details about their flight, including lateral and vertical position data (i.e. latitude, longitude, and altitude). Ground-based and satellite-based receivers can pick these signals up and store them.
This project has been underpinned by 32 full days worth of such data, spread across 2021, 2022, and up to summer 2023.Methodology
Starting with a numerical 2D array of all zeroes, we then proceed for each flight by drawing line segments between position reports. For every 'cell' that the line passes through its value is incremented by one.
After processing all 32 days' worth of flights, the result is an array that represents the 'density' of air traffic. Cells that have been flown through more frequently will have very large counts recorded (generally on the ground near airports - when aircraft don't move very far, they keep 'adding' to the cell), and those that have been flown through less frequently (or never at all) will have low or zero counts.
With the intensive processing out of the way, the final step is to turn the array into an image, by applying a colormap to the data so that we can differentiate between busier and less busy airspace, enabling us to see the traffic flows.
Applying this to all of the data gives us an initial picture. However, the dynamics of the global airspace system vary depending on phase of flight. Therefore, the data is filtered on four flight level groups, in order to allow the flows at those levels to be observed in isolation. The lowest flight level is group is surface to FL100 (that is 10,000ft above standard pressure datum, on average equivalent to 'sea level'), and captures approach and initial departure phases (for the majority of airports, other than those at high altitude). FL100 to FL240 is the next group, which shows transition to/from cruise. FL240 to FL300 captures turboprop and regional jet cruise and larger jet transition. FL300 and above captures medium and wide-body jet cruise.
More Art than Science
Whilst a lot of robust technical processing has taken place to generate the images, the end result is more art than science. Please bear this in mind when using the heatmap! Caveats include:
- The data has many 'gaps' in coverage. Just because a region looks 'empty' doesn't necessarily mean that it is not busy airspace.
- 32 days' worth of data can only tell so much. Airports change runway direction and configuration based on the wind direction. The heatmap might not capture a less frequently occuring configuration. Or it may have captured a variety of popular configurations, but due to chance, the distribution sampled may be different to how it 'typically is'.
- The colorscale used has been algorithmically tuned to allow maximum visual differentiation of traffic flows. The relationship between pixel intensity and traffic density is non-linear.
- Things change in the airspace system - aircraft fleet, flight origin/destination, airspace structure, may all change either daily, weekly, seasonally, annually, or permanently. So what can be seen through this visualisation isn't necessarily a reliable indication of what will come in the future.
So, in short: please, please, don't jump to any conclusions when using this site, and don't make any important decisions (like where to buy a house...) using this site. What you see is more akin to technical art, meant for education's sake, than science upon which to base decisions.
If you do wish to look at traffic in more detail, there are many resources online. The following are good starting points:
ADS-B Exchange - live ground-based open source ADS-B
FlightRadar24 - live ground and satellite based ADS-B
SkyVector - airspace volumes/structures
Analysis Stack
Polars for rapid, parallel and memory efficient data processing.
Numpy for scientific computing.
Folium for a flexible web map interface.
Python programming language.
Google Colaboratory for easy research-oriented cloud compute.
AWS for robust hosting with a low-latency global content distribution network.
The Future
I had a lot of fun with this project. I hope you have fun with it too!
I'm thinking of making my code, and the map tiles, open source in the future, when I get the time. Maybe that is a next step.
If you've enjoyed the site and are feeling generous, this project has mostly been fuelled by coffee, so feel free to:
If you've got any feedback, questions, or comments on the site, drop me a line at aircraft heatmap (all one word) at gmail dot com.
Thanks for dropping by!