Visualizing Flight Paths with Python

This is an example of plotting flight paths from the OpenFlights dataset with the gcmap Python package.

The OpenFlights dataset contains three files, airports.dat, routes.dat, and airlines.dat. We are interested in the first two. We'll use airports.dat to get the coordinates of each airport, and routes to determine which airports are paired by a route.

First, we use pandas to load the data.

In [1]:
import pandas as pd
import numpy as np

ROUTE_COLS = ('airline_name', 'airline_id', 'source_code', 'source_id', 'dest_code', 'dest_id', 'codeshare', 'stops', 'equiptment')
AIRPORT_COLS = ('airport_id', 'airport_name', 'city', 'country', 'iata', 'icao', 'latitude', 'longitude', 'altitude', 'timezone', 'dst')

routes = pd.read_csv('routes.dat', header=None, names=ROUTE_COLS, na_values=['\\N'])
airports = pd.read_csv('airports.dat', header=None, names=AIRPORT_COLS)

There may be multiple routes between two airports. Here we combine them into one row per pair, with a column cnt counting the number of times each pair appears.

In [2]:
airport_pairs = routes.groupby(('source_id', 'dest_id')).size()
airport_pairs = airport_pairs.reset_index()
airport_pairs.columns = ('source_id', 'dest_id', 'cnt')

Merge the airport_pairs table with the airports table to get the geocoordinates for each airport.

In [3]:
airport_pairs = airport_pairs.merge(airports, left_on='source_id', right_on='airport_id') \
                             .merge(airports, left_on='dest_id', right_on='airport_id', suffixes=('_source', '_dest'))

Now, bring in GCMapper from the gcmap package to render the visualization.

In [4]:
from gcmap import GCMapper
gcm = GCMapper()
gcm.set_data(airport_pairs.longitude_source, airport_pairs.latitude_source, airport_pairs.longitude_dest, airport_pairs.latitude_dest, airport_pairs.cnt)
img = gcm.draw()

The result is a Python Image Library image, which we can save as a png (among a handful of other formats).

In [5]:'test.png')
In [6]:
from IPython.core import display