howmanypeoplearearound

Count the number of people around you 👨‍👨‍👦 by monitoring wifi signals 📡.

howmanypeoplearearound calculates the number of people in the vicinity using the approximate number of smartphones as a proxy (since ~70% of people have smartphones nowadays). A cellphone is determined to be in proximity to the computer based on sniffing WiFi probe requests. Possible uses of howmanypeoplearearound include: monitoring foot traffic in your house with Raspberry Pis, seeing if your roommates are home, etc.

Tested on Linux (Raspbian and Ubuntu) and macOS.

It may be illegal to monitor networks for MAC addresses, especially on networks that you do not own. Please check your country's laws (for US Section 18 U.S. Code § 2511) - discussion.

Getting started

Dependencies

WiFi adapter with monitor mode

There are a number of possible USB WiFi adapters that support monitor mode. Personally I prefer the TN722N which is only ~$10 and works great with every model of the Raspberry Pi. Here is a good list of adapters that support 'ad-hoc' mode for the Raspberry Pi.

OSX / Mac

  brew install wireshark
  brew cask install wireshark-chmodbpf

Linux tshark

sudo apt-get install tshark

Then update it so it can be run as non-root:

sudo dpkg-reconfigure wireshark-common     (select YES)
sudo usermod -a -G wireshark $USER

You will need to logout and log back in for changes to effect.

Install

If you have Python installed, run this command

pip install howmanypeoplearearound

Run

Quickstart

To run, simply type in

$ howmanypeoplearearound
Using wlan1 adapter and scanning for 60 seconds...
[==================================================] 100%        0s left
There are about 3 people around.

You will be prompted for the WiFi adapter to use for scanning. Make sure to use an adapter that supports "monitor" mode.

Options

You can modify the scan time, designate the adapter, or modify the output using some command-line options.

$ howmanypeoplearearound --help

Options:
  -a, --adapter TEXT   adapter to use
  -z, --analyze TEXT   analyze file
  -s, --scantime TEXT  time in seconds to scan
  -o, --out TEXT       output cellphone data to file
  -v, --verbose        verbose mode
  --number             just print the number
  -j, --jsonprint      print JSON of cellphone data
  -n, --nearby         only quantify signals that are nearby (rssi > -70)
  --nocorrection       do not apply correction
  --loop               loop forever

Print JSON

You can generate an JSON-formatted output to see what kind of phones are around:

$ howmanypeoplearearound -o test.json -a wlan1
[==================================================] 100%         0s left
There are about 4 people around.
$ cat test.json | python3 -m json.tool
[
  {
    "rssi": -86.0,
    "mac": "90:e7:c4:xx:xx:xx",
    "company": "HTC Corporation"
  },
  {
    "rssi": -84.0,
    "mac": "80:e6:50:xx:xx:xx",
    "company": "Apple, Inc."
  },
  {
    "rssi": -49.0,
    "mac": "ac:37:43:xx:xx:xx",
    "company": "HTC Corporation"
  }
]

A higher rssi means closer (one of these phones is mine, and the other two are my roommates' who were upstairs).

Run forever

You can add --loop to make this run forever and append new lines an output file, test.json:

$ howmanypeoplearearound -o test.json -a wlan1 --loop

Visualize

You can visualize the output from a looped command via a browser using:

$ howmanypeoplearearound --analyze test.json 
Wrote index.html
Open browser to http://localhost:8001
Type Ctl+C to exit

Then just open up index.html in a browser and you should see plots. The first plot shows the number of people over time. Here you can see that people start arriving at work place around 8-9am (when work starts!).

The second plot shows the RSSI values for the mac addresses seen. You can double-click on one of them in particular to highlight that trajectory, as I have done here for my phone (you can see when I leave from and when I arrive to work!):

How does it work?

howmanypeoplearearound counts up the number of probe requests coming from cellphones in a given amount of time. The probe requests can be "sniffed" from a monitor-mode enabled WiFi adapter using tshark. An acccurate count does depend on everyone having cellphone and also scanning long enough (1 - 10 minutes) to capture the packet when a phone pings the WiFi network (which happens every 1 to 10 minutes unless the phone is off or WiFi is disabled).

This is a simplification of another program I wrote, find-lf which uses a similar idea with a cluster of Raspberry Pis to geolocate positions of cellphones within the vicinity.

License

MIT