Introduction

Feed Visualizer is a tool that can cluster RSS/Atom feed items based on semantic similarity and generate interactive visualization. This tool can be used to generate 'semantic summary' of any website by reading it's RSS/Atom feed. Shown below is an image of how the visualization generated by Feed Visualizer looks like. If you like this tool please consider giving a ⭐ on github !

Interactive Demos:

• Visualization of around 950 items from Slashdot’s RSS feed:
  📈https://ashishware.com/static/slashdot_viz.html

• Visualization of NASA’s RSS feed:
  📈https://ashishware.com/static/nasa_viz.html

• Visualization of Martin Fowler's Atom feed:
  📈https://ashishware.com/static/martin_fowler_viz.html

• Visualization of BCC's RSS feed:
  📈https://ashishware.com/static/bbc_viz.html

Quick Start

Clone the repo

git clone https://github.com/code2k13/feed-visualizer.git

Navigate to the the newly created directory

cd feed-visualizer

Install the required modules

pip install -r requirements.txt

Typically a RSS or Atom file only contains recent information from the website. This is where, I would highly recommend using wayback_machine_downloader tool. Follow the instructions on this page to install the tool.

The below command downloads public RSS feed from NASA for last few months and saves to folder named 'nasa'

wayback_machine_downloader https://www.nasa.gov/rss/dyn/breaking_news.rss -s -f 202101 -t 202106  -d nasa 

Alternatively you can simply create a new folder and paste all RSS or Atom files in it (if you have them) ! Make sure to point your config to this folder (read next step)

Now, we need to create a config file for Feed Visualizer. The config file contains path to input directory, name of output directory and some other settings (discussed later) that control the output of the tool. This is what a sample configuration file looks like :

{
    "input_directory": "nasa",
    "output_directory": "nasa_output",
    "pretrained_model": "all-mpnet-base-v2",
    "clust_dist_threshold":1,
    "tsne_iter": 8000,
    "text_max_length": 2048,
    "random_state": 45,
    "topic_str_min_df": 0.20
}

Now its time to run our tool

python3 visualize.py -c config.json

Once the above command completes, you should see visualization.html and data.csv files in the output folder (nasa_output). Copy these files to a webserver (or use a dummy server like http-server ) and view the visualization.html page in a browser. You should see something like this:

Config settings

Here is some information on what each config setting does:

{
    "input_directory": "path to input directory. Can contain subfolders. But should only contain RSS  or Atom files",
    "output_directory": "path to output directory where visualization will be stored. Directory is created if not present. Contents are always overwritten.",
    "pretrained_model": "name of pretrained model. Here is list of all valid model names https://www.sbert.net/docs/pretrained_models.html#model-overview",
    "clust_dist_threshold": "Integer representing maximum radius of cluster. There is no correct value here. Experiment !",
    "tsne_iter": "Integer representing number of iterations for TSNE (higher is better)",
    "text_max_length": "Integer representing number of characters to read from content/description for semantic encoding.",
    "random_state": "A integer to which serves as random seed while generating visualization. Use same random_state for reproducible results with set of data",
    "topic_str_min_df": "A float. For example value of 0.25 means that only phrases which are present in 25% or more items in a cluster will be considered for being used as name of the cluster."  
}

Issues/Feature Requests/Bugs

You can reach out to me on 👨‍💼 LinkedIn and 🗨️Twitter for reporting any issues/bugs or for feature requests !