A Python interface to the OpenSky Network Impala shell
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


Python interface to the OpenSky Network Impala shell

Xavier Olive, 2018
MIT license

No longer maintained: this repository is no longer active.
There are bugs that will not be fixed here.
The same functionality is available in the actively developped traffic Python library.


git clone git+https://github.com/xoolive/opensky_data
cd opensky_data
pip install -r requirements


A basic script is given as an example call to the API.

Edit the settings.cfg file with your credentials for the Impala shell.

usage: opensky_data.py [-h] [-u UNTIL] [-o OUTPUT_FILE] [-c CALLSIGN]
                       [-b BOUNDS] [-s SETTINGS]

Get data from OpenSky records

positional arguments:
  date            start date for trajectories

optional arguments:
  -h, --help      show this help message and exit
  -u UNTIL        end date for trajectories (default: date + 1 day)
  -o OUTPUT_FILE  output file for trajectories (default: output.csv)
  -c CALLSIGN     callsign for one flight
  -b BOUNDS       bounding box for trajectories (location name)
  -s SETTINGS     setting file with login information

Get flight DLH66N on November 23rd 2017 (OpenSky Workshop day!):

python opensky_data.py 2017-11-23 -c DLH66N -o DLH66N.csv

Get all trajectories over (the bounding box of) Switzerland between 6am and 7am (UTC) on January 1st:

python opensky_data.py 2018-01-01T06:00 -u 2018-01-01T07:00 -b Switzerland

See an example of the API usage in notebook.ipynb.

Under the hood

Each request is split hour by hour (see indexing issues on Impala page) and put in cache in the default temporary directory of your OS. After each iteration, a pandas dataframe is created from a modified version of the cached data (robust to network issues...): all dataframes are then concatenated and exported based on the extension of the output file.


For now, only the trajectory history use case has been addressed. Feel free to contributed with a pull request if you see any contribution that can be useful to the community.