2

I want to extract as many numerical features as I can from GPS data (longitude and latitude). I am using pandas with python. The main features I am interested in are speed, lateral and longitudinal acceleration.

a sample of the data is as below with a reading every 1 second. I searched google for an answer, however, I can only find equations for speed and map visualizations. However, I believe more features can be extracted from this data.

Longitude   Latitude
35.48396    33.902283
35.484046   33.90227
35.484128   33.902241
35.484222   33.90216
35.484282   33.90218
35.484336   33.902201
35.484431   33.902166
35.484504   33.902213
35.484582   33.9022
cs95
  • 379,657
  • 97
  • 704
  • 746
Busy Bee
  • 107
  • 2
  • 8
  • Use ML and do it yourself. – cs95 Sep 27 '17 at 11:28
  • extracting features from GPS is for ML purposes. I believe extracting lateral and longitudinal acceleration from the data would produce a better prediction. – Busy Bee Sep 27 '17 at 11:34
  • 1
    Try `df.describe`, and start working your way from there. There are many data crunching functions, such as mean, std, and so on, which you can combine with rolling windows or resampled groups. Make a genuine effort to solve this yourself first. Like I mentioned, this is not a code sweatshop. – cs95 Sep 27 '17 at 12:05

1 Answers1

2

Look at https://github.com/mrJean1/PyGeodesy

from pygeodesy.sphericalNvector import LatLon

p = LatLon(35.48396, 33.902283)
q = LatLon(35.484046, 33.90227)
d = p.distanceTo(q) # 9.64 km
K. T. Kirk
  • 21
  • 2
  • Thanks, an interesting library for distance. However, i am not sure if it is possible to extract lateral and longitudinal acceleration from the data – Busy Bee Sep 28 '17 at 10:58
  • Just calculate some Velocity between locations then you can calculate acceleration. http://www.physicsclassroom.com/class/1DKin/Lesson-1/Acceleration – K. T. Kirk Oct 04 '17 at 17:04