How to access sample trajectories?

A bundle of sample trajectories covering various situations has been included in the library for reference, testing and for providing a baseline to compare the performance of various methods and algorithms.

All sample trajectories are available in the traffic.data.samples module. The import automatically dispatch to Flight or Traffic according to the nature of the data.

Note

A subset of the sample trajectories are presented on this page. Other parts of the documentation (e.g. calibration flights or trajectory clustering) may refer to other available sample trajectories.

Belevingsvlucht

https://www.belevingsvlucht.nl/

On May 30th 2018, test flights were conducted along future routes arriving and departing from Lelystad Airport in the Netherlands. The purpose of this flight operated by a Boeing 737 owned by Transavia was to assess noise exposure in neighbouring cities.

from traffic.data.samples import belevingsvlucht
belevingsvlucht

Flight

callsign: TRA051
aircraft: 484506 · 🇳🇱 PH-HZO (B738)
start: 2018-05-30 15:21:38+00:00
stop: 2018-05-30 20:22:56+00:00
duration: 0 days 05:01:18
sampling rate: 1 second(s)

Dreamliner Air France

After getting their first Boeing 787 Dreamliner, Air France equipped a Socata TBM 900 and shot the following video in 8k. Both trajectories are (partly) available as sample flights.

You can explore how the two trajectories intertwine for shooting purposes:

from traffic.data.samples import dreamliner_airfrance
dreamliner_airfrance

Traffic

with 2 identifiers

		count
icao24	callsign
39c424	AFR787V	13143
3900fb	FWKDL	4767

dreamliner_airfrance['AFR787V'] | dreamliner_airfrance['FWKDL']

Flight

callsign: AFR787V
aircraft: 39c424 · 🇫🇷 F-HRBE (B789)
start: 2017-12-01 12:47:57+00:00
stop: 2017-12-01 16:26:59+00:00
duration: 0 days 03:39:02
sampling rate: 1 second(s)

Flight

callsign: FWKDL
aircraft: 3900fb · 🇫🇷 F-WKDL (TBM7)
start: 2017-12-01 14:40:33+00:00
stop: 2017-12-01 15:59:59+00:00
duration: 0 days 01:19:26
sampling rate: 1 second(s)

dreamliner_airfrance.map_leaflet(
    highlight=dict(red=lambda f: f.query('icao24 == "3900fb"')),
    center=(43.5, 3.37),
    zoom=9,
)

The following Altair specification lets you plot the altitude profile (clamped to 20,000 ft) for the two trajectories. The steps in the altitude profile are due to the low sampling rate in the data:

import altair as alt

chart = (
    alt.layer(
        *list(
            flight.chart().encode(
                alt.X("utchoursminutes(timestamp)", title=""),
                alt.Y(
                    "altitude",
                    title="",
                    scale=alt.Scale(domain=(5000, 20000), clamp=True),
                ),
                alt.Color("callsign", title="Callsign of the aircraft"),
            )
            for flight in dreamliner_airfrance.between(
                "2017-12-01 14:40", "2017-12-01 15:40"
            )
        )
    )
    .properties(title="Altitude (in ft)", width=600)
    .configure_title(anchor="start", font="Lato", fontSize=15, dy=-5)
    .configure_axis(labelFontSize=12)
    .configure_legend(orient="bottom", titleFont="Lato", titleFontSize=13)
)
chart

Airbus tree

Before Christmas 2017, an Airbus pilot in Germany has delivered an early festive present by tracing the outline of an enormous Christmas tree during a test flight.

from traffic.data.samples import airbus_tree
airbus_tree

Flight

callsign: AIB232E
aircraft: 3807fa
start: 2017-12-13 11:47:15+00:00
stop: 2017-12-13 17:08:40+00:00
duration: 0 days 05:21:25
sampling rate: 4 second(s)

Other trajectories

Even though all trajectories are accessible from the traffic.data.samples module, they are in practice organised by categories:

from pkgutil import iter_modules  # one of Python inspection modules
from traffic.data import samples

list(category.name for category in iter_modules(samples.__path__))

['ambulances',
 'calibration',
 'collections',
 'featured',
 'firefighting',
 'fr24',
 'gliders',
 'military',
 'noisy',
 'onground',
 'performance',
 'surveillance',
 'surveys',
 'tv_relay']

For instance, National Geographic Institutes sometimes conduct aerial surveys including photography and LIDAR measurements from aircraft, some of such trajectories are available in the surveys module (category).

In each category, you can list all available trajectories:

from traffic.data.samples import surveys
surveys.__all__

['georeal_fyn_island',
 'ign_fontainebleau',
 'ign_lot_et_garonne',
 'ign_mercantour',
 'indiana',
 'liguria',
 'mecsek_mountains',
 'pixair_toulouse',
 'texas']

Then the same trajectory is available both from the module of each category (here surveys) and from the root samples module:

surveys.pixair_toulouse | samples.pixair_toulouse

Flight

callsign: PXR31F
aircraft: 39b861 · 🇫🇷 F-HODB (PA31)
start: 2019-03-31 08:52:04+00:00
stop: 2019-03-31 13:32:34+00:00
duration: 0 days 04:40:30
sampling rate: -9 second(s)

Flight

callsign: PXR31F
aircraft: 39b861 · 🇫🇷 F-HODB (PA31)
start: 2019-03-31 08:52:04+00:00
stop: 2019-03-31 13:32:34+00:00
duration: 0 days 04:40:30
sampling rate: -9 second(s)

So in practice, both imports are valid (and completion helps in both cases):

from traffic.data.samples import pixair_toulouse
from traffic.data.samples.surveys import pixair_toulouse

You can make a Traffic object from a category module:

from traffic.core import Traffic

t_surveys = Traffic.from_flights(  # actually, this is equivalent to sum(...)
    getattr(surveys, name)  # getattr(surveys, "pixair_toulouse") is surveys.pixair_toulouse
    .assign(flight_id=name)  # gives a flight_id to each trajectory
    for name in surveys.__all__
)
t_surveys

Traffic

with 9 identifiers

	count
flight_id
ign_fontainebleau	1930
pixair_toulouse	1819
liguria	1516
mecsek_mountains	1462
georeal_fyn_island	1254
ign_lot_et_garonne	814
indiana	755
texas	703
ign_mercantour	630

As illustrated throughout this documentation, the | operator (or_) concatenates Jupyter representations of eligible objects. Therefore, the following trick lets you apply the or_ operator to all flights available in the surveys category:

from functools import reduce
from operator import or_

# this will do surveys.flight1 | surveys.flight2 | surveys.flight3 | etc.
reduce(or_, t_surveys)

Flight georeal_fyn_island

callsign: OKVOK
aircraft: 49d1d3 · 🇨🇿 OK-VOK (C421)
start: 2018-04-19 07:31:00+00:00
stop: 2018-04-19 10:59:50+00:00
duration: 0 days 03:28:50
sampling rate: 10 second(s)

Flight ign_fontainebleau

callsign: FGALN
aircraft: 39016d · 🇫🇷 F-GALN (BE20)
start: 2019-02-27 08:51:13+00:00
stop: 2019-02-27 16:04:21+00:00
duration: 0 days 07:13:08
sampling rate: 13 second(s)

Flight ign_lot_et_garonne

callsign: FGALN
aircraft: 39016d · 🇫🇷 F-GALN (BE20)
start: 2017-06-10 07:04:10+00:00
stop: 2017-06-10 12:09:48+00:00
duration: 0 days 05:05:38
sampling rate: 23 second(s)

Flight ign_mercantour

callsign: FGALN
aircraft: 39016d · 🇫🇷 F-GALN (BE20)
start: 2017-07-06 06:53:36+00:00
stop: 2017-07-06 10:28:31+00:00
duration: 0 days 03:34:55
sampling rate: 21 second(s)

Flight indiana

callsign: N441FS
aircraft: a54ee2 · 🇺🇸 N441FS (C441)
start: 2018-07-08 13:03:44+00:00
stop: 2018-07-08 18:12:26+00:00
duration: 0 days 05:08:42
sampling rate: 25 second(s)

Flight liguria

callsign: IGRAD
aircraft: 3006bd · 🇮🇹 I-GRAD (P68)
start: 2020-04-15 07:39:00+00:00
stop: 2020-04-15 11:51:30+00:00
duration: 0 days 04:12:30
sampling rate: 10 second(s)

Flight mecsek_mountains

callsign: HAYCG
aircraft: 473f9a · 🇭🇺 HA-YCG (PA27)
start: 2020-03-17 11:40:10+00:00
stop: 2020-03-17 15:43:40+00:00
duration: 0 days 04:03:30
sampling rate: 10 second(s)

Flight pixair_toulouse

callsign: PXR31F
aircraft: 39b861 · 🇫🇷 F-HODB (PA31)
start: 2019-03-31 08:52:04+00:00
stop: 2019-03-31 13:32:34+00:00
duration: 0 days 04:40:30
sampling rate: -9 second(s)

Flight texas

callsign: N441FS
aircraft: a54ee2 · 🇺🇸 N441FS (C441)
start: 2018-10-27 14:38:40+00:00
stop: 2018-10-27 19:13:55+00:00
duration: 0 days 04:35:15
sampling rate: 24 second(s)