Contains executables for scraping and processing webcam data from the internet. Also contains executables for analyzing these webcams. Currently, we support semantic clustering of webcams.

cluster/

This directory ontains source code for semantic clustering. To build, run the following from the cluster/ directory:

mkdir build
cd build
cmake ..
make

To run, run the following from the analysis/ directory:

./COMPLETE <frames_dir>

Currently, you must recompile the program everytime you would like to experiment with different hyperparameters. Feel free to modify cluster/src/complete.cc to accept command-line arguments instead.

scrape_metadata.py

Scrapes metadata associated with webcams from sources. To use this executable, specify the following:

• source: The name of the source of the webcams.
• identifiers: A range of unique identifiers for the webcams.

scrape_frames.py

Scrapes frames associated with webcams from sources. To use this executable, specifiy the following:

• source: The name of the source of the webcams.
• identifiers: A range of unique identifiers for the webcams.
• period: The period (in minutes) between each frame.
• threads: The number of scraping threads to use.

list_metadata.py

Lists metadata associated with scraped webcams. You may specify the following flag:

• live: Only shows metadata for live webcams.