MISCELA_demo

This project builds a DEMO system for spatio-temporal sensor data mining, MISCELA-V.

requirements

• docker
• docker-compose

quick start

1. Run api server by docker-compose.

cd backend
bash run_containers.sh

API server on django and MongoDB start on your local environment. You can check all API urls in ./backend/miscela_api/api/urls.py and main view file is ./backend/miscela_api/api/views.py. All native MISCELA programs are under ./backend/miscela_api/api/src/ Console outputs are redirected to ./backend/miscela_log.txt

2. Open index.html in your web-browser.

You can upload our datasets at Your Dataset page via a user interface that provides two ways of data upload: drag-and-drop and selecting files from finder. For uploading datasets, we need to prepare three files; data.csv, location.csv, and attribute.csv. data.csv.

The datasets folder contains the sample datasets. You can check the format of the dataset files and try to run them easily.

3. Set parameters and run MISCELA.

You can run MISCELA by setting parameters Max Attribute, Minimum Support, Evolving rate and Distance and then pushing the Update button.

4. Draw a graph of the set of CAP sensors.

After the calculation is completed, the CAP sensor set is displayed on the map. You select a sensor and click on it to discover correlations between multiple attributes from a set of spatially proximate sensors, and to check that the measurements are temporally correlated.