TKM Python implementation

TKM stands for Threshold-based k-means Monitoring

This is a Python implementation of paper Continuous k-Means Monitoring over Moving Objects by Zhenjie Zhang, Yin Yang, Anthony K.H. Tung, and Dimitris Papadias. This repisotory consists of three components:

Notebooks (under root folder)

modified_hill_climbing_demo.ipynb: known as "HC*" in the original paper
threshold_demo.ipynb: implemented the Computation of thresholds section of the paper
process_taxi_data.ipynb: processed Microsoft's T-Drive trajectory data sample into a JSON file that's readable by the previous two algorithm implementations and the web app.

Web app (under `flaskr` folder)

This implementation includes a simple web app powered by Google Maps JavaScript API that visualizes JSON outputs in the format specified in process_taxi_data.ipynb. All you need to do is to place the generated .json file under \flaskr\static folder and change the filename param of $.getJSON("{{ url_for('static', filename='sample_1.json') }}", function(res) { ... } in the \flaskr\templates\index.html to the name of your json file.

Before running the app on your own

You need to replace api_key with your own Google Maps API key in \flaskr\__init__.py. Then run the app in dev mode using:

export FLASK_APP=flaskr
export FLASK_ENV=development
flask run

Mobile app (under `app` folder)

This implementation also includes a simple Flutter app that automatically updates the device's location according to the received threshold.

This demo app assumes these API URLs:

/api/init_location, a POST API with JSON post data:

{
    "id": "0",
    "lat": 37.421998,
    "lng": -122.084
}

if succeeded, the server should respond one int number indicating its updated threshold (in meters).

/api/update_location, a POST API with JSON post data:

{
    "id": "0",
    "lat": 37.451998,
    "lng": -122.094
}

if succeeded, the server should respond one int number indicating its updated threshold (in meters).

Before compiling the app on your own

You need to 1) set up Firebase in the app using your own Firebase credentials, 2) change API_BASE_URL and SENSITIVITY two constants under \app\lib\main.dart accordingly.

For information about how to configure Firebase Cloud Messaging for this app, please refer to this documentation on the FCM Flutter plugin.

For information about how to send a message from the server to any specific mobile device or a group of devices, please refer to this sample implementation of mine.

le0tan/python-threshold-based-k-means-monitoring