I reconstructed my dataset in pandas
DataFrame
by using a multi-index
, and it is now in the following format.
In [1]: df.head(12)
Out [1]:
In order to put it into a GeoJSON
LineString
format and visualize it on a map, I need to write a Python
loop
over each point and each line through millions of satellite observational points. For reference, the following example specifies a GeoJSON
LineString
.
{ type: "LineString", coordinates: [ [ 40, 5 ], [ 41, 6 ] ] }
However, not always as shown in the figure that a line consists of 4 points for the first three lines, the number of points for a specific line in this dataset is totally random, ranging from 4 to hundreds.
I am so confused how to write a Python
loop
that could help me put my coordinates into GeoJSON
LineString
type by using a multi-index
, e.g.
In [2]: df.Longitude[1][4]
Out [2]: 128
Thanks for your time!
A combination of groupby
and to_json
seems to work well.
import pandas as pd
import numpy as np
import pprint
arrays = [np.array([1, 1, 1, 2, 2, 3, 3, 3, 3, 3, 4, 4]),
np.array([1, 2, 3, 1, 2, 1, 2, 3, 4, 5, 1, 2])]
df = pd.DataFrame(np.arange(24).reshape(12,2),
index=arrays, columns=['Longitude', 'Lattitude'])
dd = {"type":"Feature",
"geometry":{"type":"Linestring",
"coordinates":None
},
"properties":{"prop0":'red',
"prop1":'dashed'
}
}
for _, group in df.groupby(level=0):
dd["geometry"]["coordinates"] = group.to_json(orient='values')
pprint.pprint(dd)
output:
{'geometry': {'coordinates': '[[0,1],[2,3],[4,5]]',
'type': 'Linestring'},
'properties': {'prop0': 'red',
'prop1': 'dashed'},
'type': 'Feature'}
{'geometry': {'coordinates': '[[6,7],[8,9]]',
'type': 'Linestring'},
'properties': {'prop0': 'red',
'prop1': 'dashed'},
'type': 'Feature'}
{'geometry': {'coordinates': '[[10,11],[12,13],[14,15],[16,17],[18,19]]',
'type': 'Linestring'},
'properties': {'prop0': 'red',
'prop1': 'dashed'},
'type': 'Feature'}
{'geometry': {'coordinates': '[[20,21],[22,23]]',
'type': 'Linestring'},
'properties': {'prop0': 'red',
'prop1': 'dashed'},
'type': 'Feature'}