itinerum-tripkit 0.0.26

A toolkit for inferencing trips and trip metadata from Itinerum GPS data

itinerum-tripkit

Documentation for library usage: https://itinerum-tripkit.readthedocs.io/

This library serves as a framework to process data from the Itinerum platform and hardware GPS loggers (e.g., QStarz). It can be used both through as a library in Jupyter to explore datasets interactively or imported as a module in standalone scripts and applications.

This repository also serves as the development bed for the Itinerum platform algorithms within the TRIP Lab repositories.

Looking to get started without coding? Try the itinerum-tripkit-cli!

Setup

Quickstart

1. Install this library from PyPI (a Python virtual environment is recommended)
2. Create a configuration file with input filepaths, output filepaths, and trip processing parameters. See the included tripkit_config.py file for a full example.
3. Import tripkit as a dependency in a notebook or script

For more complete installation information, see the official itinerum-tripkit documentation.

Loading Subway Stations

Subway station data for trip detection can be loaded similarly for all processing modules. Place a .csv file of station entrances with the columns of x (or longitude) and y (or latitude). Locations are expected as geographic coordinates only. Edit the SUBWAY_STATIONS_FP config parameter to reflect the subway stations .csv filepath.

Example

View attributes on a User

import tripkit_config
itinerum = Itinerum(tripkit_config)

# create a new database and read in .csv data
itinerum.setup()

# load all users from database
users = itinerum.load_all_users()

test_user = users[0]
print(test_user.coordinates)
print(test_user.prompt_responses)

Run trip detection on a User

import tripkit_config
itinerum = Itinerum(tripkit_config)

# load user from database by uuid
user = itinerum.database.load_user('00000000-0000-0000-0000-000000000000')

# run a provided trip detection algorithm
parameters = {
    'subway_entrances': itinerum.database.load_subway_entrances(),
    'break_interval_seconds': tripkit_config.TRIP_DETECTION_BREAK_INTERVAL_SECONDS,
    'subway_buffer_meters': tripkit_config.TRIP_DETECTION_SUBWAY_BUFFER_METERS,
    'cold_start_distance': tripkit_config.TRIP_DETECTION_COLD_START_DISTANCE_METERS,
    'accuracy_cutoff_meters': tripkit_config.TRIP_DETECTION_ACCURACY_CUTOFF_METERS
}
trips = itinerum.process.trip_detection.triplab.v2.algorithm.run(user.coordinates, parameters)

Processing

Trip Detection

Arguments
parameters	A dictionary to supply arbitrary kwargs to an algorithm
subway_stations	A list of subway station entrance database objects containing latitude and longitude attributes
coordinates	A timestamp-ordered list of coordinates as dicts for a specific user. Multiple users should be run in sequence and have their output coordinates concatenated into a single list after if desired.

Trip Outputs

Trips will be output with the following trip codes to indicate the type of trip:

Trip Code	Description
1	Complete trip
2	Complete trip - subway
101	Missing trip
102	Missing trip - subway
103	Missing trip - less than 250m
201	Single point
202	Distance too short - less than 250m

Outputs

The aim of this library is to provide easy visualization of Itinerum data to assist in writing trip processing algorthms. Therefore at a minimum, the library provides exporting processed coordinates and traces as .geojson files (TBA: GeoPackage format). With a PostgreSQL backend for caching, PostGIS can be enabled (unimplemented) and a geom column generated for directly connection QGIS to the output data. The library should also easily provide methods for easily plotting GPS within Jupyter notebooks.