Edge Rank

Python implementation of the EdgeRank algorithm for ranking statuses in the news feed of social networks.

Overview

On application start, you can login just by entering a name present in the dataset (try "Sarina Hudgens" or "Angel Harmison").

Then, 10 most relevant statuses will be displayed. These are the statuses with the highest EdgeRank score, which is calculated as follows:

EdgeRank = user_affinity * status_weight * time_decay_factor

where:

user_affinity is the affinity of the user to the author of the status
status_weight is determined by the statuses popularity (number of likes, comments and shares)
time_decay_factor is a factor that decreases the score of older statuses

Then, the status search is shown. The search looks for statuses that contain the given keywords in the message, with statuses with higher EdgeRank score being displayed first.

Running

To run the program, simply start main.py with python3

python3 main.py

Loading data

If there are no serialized data structures, the program will try to load them from the dataset. You can start this operation manually by running the load_data.py script.

python3 load_data.py

A new dataset can be appended to existing data by running the append_data.py script.

python3 append_data.py

The test files can be updated to use newer dates by running the following script.

python3 data_utils/parse_files.py