bw_stats is a front-end webapp to visualize data gathered by it's own python script from OpenWRT/LEDE library wrtbwmon
A properly configured wrtbwmon tool running on a router produces a usage.db file that can be published by a symlink to the /www/ folder and can then be parsed by an external tool.
bw_stats can be run on the router itself, therefore having no need to publish the file but, in my case, I'm running bw_stats on a small headless server running on an orange pi so the file needs to be fetched remotely.
By parsing the usage.db file (which is in fact a CSV file) every 5 seconds, bw_stats builds a one month long database. The front-end app then uses this database to show nice graphics using your bandwith-usage data.
The actual tools and visualizations (wrtbwmon included) are very limited in terms of information. The Luci visualization for wrtbwmon shows no graphs but a improved version of the usage table that wrtbwmon itself provides, on the other hand, tools like nlbwmon that actually have graphs have a better approach and allow to select accounting period for them, but no tool/visualization includes realtime bandwith usage graphs.
After setting up wrtbwmon and scheduling the task to publish the usage every 5 seconds (luci-app-wrtbwmon already does this for you, so you don't have to do it) and making usage.db available with a symlink to /www/, just copy the latest release of bw_stats to your router or server and make it available from any html server running within it.
Modify the db/log_usage.py file variables to set the stats.json and usage.db locations.
Set the db/log_usage.py script to run on server boot, remember to run it on second plane to don't hang up your server forever. The script loops itself so you don't have to schedule a cron task for it.
With everything set up, just point any browser to the url where bw_stats is available and enjoy your cool graphs.