A cross-platform desktop application for collecting and visualizing WiFi signal strength across a physical space. Supports Windows, Linux, and macOS. Works with or without a floorplan image — though providing one significantly increases the accuracy and usefulness of the resulting heatmap.

🔌 Requirements 🔌

Python 3.10+
tkinter (usually bundled with Python; on Linux: sudo apt install python3-tk)

⬇️ Installation ⬇️

pip install -r requirements.txt

Dependencies: numpy, scipy, matplotlib, Pillow

No third-party WiFi libraries are required. The scanner uses native OS APIs directly (wlanapi on Windows, iw / nmcli on Linux, airport on macOS).

📖 Usage 📖

# Launch the GUI
python main.py

# Load an existing session on startup
python main.py --session my_office.json

🗺️ Interface Overview 🗺️

The application has two tabs and a persistent left sidebar.

Sidebar (always visible)

Session management — New, Open, Save, Save As
Floorplan — Load or remove a floorplan image. A note is shown when none is loaded reminding you that providing one increases accuracy.
Export PNG — Save the current heatmap view as an image file.
Session info — Shows the current session name, file path, and mode.

Tab 1 — Access Points

The first tab you see on launch. Shows a scrollable table of every access point ever discovered during the session — including those that have gone out of range, which remain in the list permanently and are shown dimmed with an "out of range" indicator in the Level column.

Table columns:

Column	Description
✓	Checkbox to include this BSSID in heatmap rendering
SSID	Network name
BSSID	Hardware MAC address — unique per physical radio
Ch	802.11 channel number
Frequency	Centre frequency in MHz
Ch Width	Channel width (20 / 40 / 80 / 160 MHz)
Band	2.4 GHz or 5 GHz
Security	WPA3-Personal, WPA2-Personal, WPA2-Enterprise, WPA-Personal, WEP, Open
Vendor	Manufacturer derived from BSSID OUI prefix (~100 vendors recognized)
Mode	802.11 mode: ax (Wi-Fi 6), ac (Wi-Fi 5), n (Wi-Fi 4), g, a, b
Level	Color-coded signal bar (green ≥ -50, yellow -50–-65, orange -65–-75, red < -75)
Last Seen	How long ago this BSSID was last visible (updates each scan)
Max	Strongest signal ever recorded for this BSSID across all scans
Min	Weakest signal ever recorded
Avg	Running average of all valid readings

Toolbar controls:

🔄 Refresh Scan — Trigger a manual scan immediately.
☑ Select All / ☐ Deselect All — Check or uncheck all rows for heatmap use.
Auto-scan — Enable automatic periodic scanning. The interval is user-configurable via a spinbox (minimum 3 seconds, default 10 seconds, maximum 300 seconds). A live countdown shows time until the next scan. While auto-scan is active the manual Refresh button is disabled to prevent scan conflicts.

The table scrolls both horizontally and vertically. Column headers are pixel-exact aligned to their data columns.

Tab 2 — Heatmap

The heatmap canvas with measurement tools.

Active Networks panel (right side):

Shows which BSSIDs are currently selected for rendering, pulled directly from the checkboxes on the Access Points tab. No separate dropdown is needed — the AP tab checkboxes are the sole selector.

Single BSSID selected → heatmap shows that radio's individual coverage.
Multiple BSSIDs selected → signal values are averaged per measurement point. Only BSSIDs that were actually visible at each location contribute to the average — a radio that was out of range at a given spot is excluded rather than dragging the average down. This is the correct way to map mesh networks.

A mode label shows either "Mode: single BSSID" or "Mode: averaging N BSSIDs".

Measurements panel (right side, full height):

A scrollable list of every recorded measurement point showing its canvas coordinates and how many APs were visible at that location. Individual measurements can be deleted by selecting them and clicking 🗑 Delete Selected. The Undo button removes the most recent point.

Heatmap rendering:

The heatmap is not drawn until at least 4 measurement points are collected for the selected BSSID(s). Until then a progress indicator shows how many points have been collected and how many more are needed.
Colored dots mark each valid measurement point, labeled with the dBm value.
Black dots with a ? mark measurement points where none of the selected BSSIDs were visible — the user was there and scanned, but that radio had no coverage at that location.
The colorbar runs from black (no signal / not found) at the bottom through navy → blue → cyan → green → yellow → orange → red (excellent) at the top. A "No signal" label appears below the scale.
When no floorplan is loaded, a subtle coordinate grid is drawn instead, and a warning note appears in the heatmap subtitle.

Workflow

Launch the app. It opens on the Access Points tab.
Click 🔄 Refresh Scan to discover nearby networks. All found networks are auto-selected on the first scan.
(Optional but recommended) Click 🖼 Load Floorplan in the sidebar. Load a PNG or JPEG overhead image of your space. This gives the interpolation real spatial context and produces a significantly more accurate heatmap.
Switch to the Heatmap tab. The Active Networks panel shows your selected BSSIDs and the current mode (single or averaged).
Walk to a location in your space and click ▶ Scan & Place. The app scans WiFi (10 samples, 100 ms apart, median reported) and enters placement mode. Click your current position on the canvas.
Repeat from step 5. Aim for at least 15–20 measurement points for a good interpolation. Prioritize:
- Locations very close to each router (expect -30 to -45 dBm)
- Far corners and edges of the space (expect -65 to -80 dBm)
- Both sides of walls and doorways
- Any spot you suspect has poor coverage
The heatmap auto-updates after each new point once the 4-point minimum is met.
Use Export PNG to save the result, or Save Session to continue later.

Mapping a mesh network accurately

If you have a mesh system (e.g. Netgear Orbi, Eero, Google Nest WiFi) where multiple radios broadcast the same SSID:

Each physical radio has a unique BSSID — use the Access Points tab to identify them by channel, frequency, or vendor.
To map one radio in isolation, check only its BSSID and uncheck the others before scanning.
Your device will silently roam between radios as you move. If all BSSIDs are checked, the multi-BSSID averaging mode will produce a combined coverage map.
For the most accurate single-radio map, consider temporarily disabling the other nodes in your mesh system's admin panel during the session.

Scanning Backends

The scanner automatically selects the best available backend per OS and falls back gracefully if the primary is unavailable.

OS	Primary	Fallback 1	Fallback 2
Windows	`wlanapi` via ctypes	`netsh wlan show networks`	—
Linux	`iw dev scan`	`nmcli dev wifi list`	`iwlist scan`
macOS	`airport -s`	—	—

Why `wlanapi` and `iw` matter 💡

Windows — wlanapi vs netsh:

netsh reports signal as a 0–100% quality score that is converted back to dBm with (pct/2) - 100. Most drivers clamp everything above roughly -55 dBm to 100%, which converts to exactly -50 dBm — meaning all strong signals appear identical and genuine variation close to the router is invisible. wlanapi calls WlanGetNetworkBssList() directly, which returns the raw dot11_RSSI field from the BSS entry struct in hardware dBm with no clamping or rounding. This can reveal 8–15 dBm of real variation that netsh completely hides. WlanScan() is also called first to request a fresh radio sweep rather than reading the OS scan cache (which can be 30–60 seconds stale).

Linux — iw vs nmcli:

nmcli reads from NetworkManager's internal scan cache and applies its own signal smoothing. iw dev <iface> scan calls the kernel's nl80211 layer directly via netlink, forcing a live hardware scan and returning signal in mBm (millibelsmilliwatt) at 0.1 dBm precision, e.g. -6500 = -65.0 dBm. It also returns raw Information Elements (IEs) from the beacon frame, enabling accurate detection of HE / VHT / HT capabilities for the Mode column.

Linux permissions for `iw`

iw dev scan requires either root or CAP_NET_RAW on the Python binary:

# Preferred: grant capability to Python (no sudo needed at runtime)
sudo setcap cap_net_raw+eip $(which python3)

# Alternative: run as root (not recommended for daily use)
sudo python main.py

If neither is available the scanner falls back to nmcli automatically.

Windows permissions for `wlanapi`

Normal user processes have WLAN_WRITE_ACCESS by default on Windows 7 and later — no elevation required. On locked-down corporate machines where this access is denied, the scanner falls back to netsh automatically.

Signal Sampling

Each ▶ Scan & Place measurement takes 10 samples at 100 ms intervals (approximately 1.1 seconds total). The median of the samples is reported, not the mean. The median discards burst spikes and interference events that can momentarily shift RSSI 10–15 dBm in either direction, producing a more stable reading that better represents the true signal at that location.

The Access Points tab Refresh Scan uses the same sampling parameters. The Auto-scan feature uses the same backend but runs on a timer rather than being placement-triggered, and accumulates running Max / Min / Avg statistics per BSSID across the entire session.

Interpolation

Signal values between measurement points are interpolated using scipy:

Points collected	Method used
< 4	No heatmap (progress indicator shown)
4–8	Cubic (Delaunay triangulation)
9+	RBF — Radial Basis Function with thin-plate spline kernel

RBF produces smoother, more physically plausible gradients for signal propagation and is preferred once enough data is available. NaN cells at the edges of the interpolation are filled with nearest-neighbor fallback so the entire canvas is always covered.

Signal Quality Reference

dBm range	Quality	Colorbar
-30 to -50	Excellent	Red → Orange → Yellow
-50 to -65	Good	Yellow → Green
-65 to -75	Fair	Green → Cyan
-75 to -90	Poor	Cyan → Blue → Navy
Not found	No signal	Black dot (●) on map

Session Files

Sessions are saved as .json files containing all measurement points, signal readings for every visible BSSID at each point, SSID mappings, the floorplan path, and canvas dimensions. Sessions can be loaded at startup:

python main.py --session my_office.json

Measurements can be deleted individually from the Heatmap tab. The session is not auto-saved — use 💾 Save Session or 💾 Save As… from the sidebar.

Floorplan Tips

Any image format works: PNG, JPEG, BMP, TIFF. PNG with transparency is supported.
The image is scaled to fit the canvas automatically — no need to resize it first.
Even a hand-drawn photo of a sketch works.
Providing a floorplan increases accuracy — measurement points map to real room features, giving the interpolation genuine spatial context.
Without a floorplan the canvas shows a coordinate grid and a warning is shown in the heatmap subtitle and export.

Project Files

File	Purpose
`main.py`	Tkinter GUI — two-tab layout (Access Points + Heatmap), sidebar, auto-scan engine, AP stats tracking
`scanner.py`	Cross-platform WiFi scanning with `wlanapi` / `iw` / `nmcli` / `airport` backends, median averaging
`interpolator.py`	Spatial interpolation — auto-selects RBF, cubic, or linear based on point count
`renderer.py`	Matplotlib heatmap rendering — colorbar, measurement dots, missing-signal black dots
`data.py`	Session model — measurements, multi-BSSID averaging, missing-point detection, JSON persistence
`requirements.txt`	Python dependencies (`numpy`, `scipy`, `matplotlib`, `Pillow`)

README.md Unescape Escape

📡 pyFi