Wigle_Sort

WiGLE CSV merger, deduplicator, region splitter, and channel analyzer for wardriving data.

---

Versions

Version	File	Description
v1	wigle_sort.py	4-region splitter (US, India, Europe, Canada)
v2	wigle_sort_v2.py	46 countries + all 50 US states/DC + channel analysis
v3	wigle_sort_v3.py	v2 + small state bounding box fix (RI, DC, etc.)
—	wigle_channel_compare.py	Side-by-side channel distribution comparison between two scan datasets

---

Requirements

pip install pandas

Python 3.8+

---

Usage

v1

python wigle_sort.py \
  --input  "E:\TECH DATA\wardrive\Wigle Full Downloads" \
  --output "E:\TECH DATA\wardrive\Wigle Sorted"

Arguments

Flag	Required	Description
--input	✅	Root folder containing WiGLE CSV files (searched recursively)
--output	✅	Output folder for sorted CSVs
--types	❌	Comma-separated types to include. Default: WIFI,BLE,BT,GSM,LTE,WCDMA

Output — v1

output/
├── wigle_us.csv
├── wigle_india.csv
├── wigle_europe.csv
├── wigle_canada.csv
├── wigle_other.csv
├── wigle_master_deduped.csv
└── wigle_no_gps.csv

---

v2, v3

python wigle_sort_v3.py \
  --input  "E:\TECH DATA\wardrive\Wigle Full Downloads" \
  --output "E:\TECH DATA\wardrive\Wigle Sorted"

Arguments

Flag	Required	Description
--input	✅	Root folder containing WiGLE CSV files (searched recursively)
--output	✅	Output folder for sorted CSVs
--types	❌	Comma-separated types to include. Default: WIFI,BLE,BT,GSM,LTE,WCDMA
--no-channel-summaries	❌	Skip channel analysis CSVs for faster runs

Output — v2/v3

output/
├── wigle_master_deduped.csv
├── wigle_no_gps.csv
├── countries/
│   ├── wigle_usa.csv
│   ├── wigle_india.csv
│   ├── wigle_netherlands.csv
│   └── wigle_<country>.csv   (46 countries total)
├── us_states/
│   ├── wigle_us_massachusetts.csv
│   ├── wigle_us_texas.csv
│   └── wigle_us_<state>.csv  (50 states + DC)
└── channel_summaries/
    ├── channels_usa.csv
    ├── channels_us_massachusetts.csv
    └── channels_<region>.csv

---

wigle_channel_compare.py

Compares channel distributions between two WiGLE scan datasets side by side. Useful for A/B testing scan configurations — for example, comparing 12 nodes on 1 device vs 6 nodes on 2 devices to see how node distribution affects 2.4 GHz vs 5 GHz coverage.

Note: WiGLE server-side CSV exports strip channel data, logging -1 for all records. Run this script against local device logs (e.g. Biscuit Ultra exports) that retain real channel values.

# Single CSV files
python wigle_channel_compare.py --a scan_a.csv --b scan_b.csv

# Whole directories (reads all CSVs recursively, same as wigle_sort)
python wigle_channel_compare.py --a ./scan_a/ --b ./scan_b/ --out compare.csv

Arguments

Flag	Required	Description
--a	✅	Scan A: CSV file or directory of WiGLE-format CSVs
--b	✅	Scan B: CSV file or directory of WiGLE-format CSVs
--label-a	❌	Label for dataset A (default: Scan_A)
--label-b	❌	Label for dataset B (default: Scan_B)
--out	❌	Save CSV report to this file

Output

Prints to terminal:

• Band summary (2.4 GHz vs 5 GHz count, percentage, and delta between scans)
• Per-channel breakdown with UNII group labels for 5 GHz channels
• Channels where the two scans diverge by ≥1% are flagged with ◄ or ▼

Optionally saves a CSV report with both tables via --out.

Deduplication

Records are deduplicated by MAC within each dataset (keeping strongest RSSI) before any channel counts are calculated, matching wigle_sort behavior. This prevents long drives inflating counts for channels seen repeatedly.

---

How it works

Input

All scripts recursively scan their input path for *.csv files. WiGLE CSV exports have a metadata header on row 1 — all scripts skip it automatically. Files with missing or malformed headers are skipped with a warning.

Deduplication

Records are deduplicated by MAC + Type. Where the same MAC appears across multiple sessions, the row with the strongest RSSI is kept.

Classification

Both versions use bounding box classification — latitude/longitude ranges per region. No reverse geocoding, no internet dependency.

• v1 — 4 coarse regions: US, India, Europe (entire continent), Canada
• v2/v3 — 46 individual countries + all 50 US states + Washington DC. US records are written to both countries/wigle_usa.csv and the appropriate us_states/ file.

Records with 0,0 coordinates, impossible values, or coordinates outside all bounding boxes (e.g. open ocean) go to wigle_other.csv.

Channel analysis (v2, v3, and wigle_channel_compare)

For every output region and state, v2/v3 generate a channel_summaries/channels_<region>.csv containing band totals and per-channel counts. wigle_channel_compare.py performs the same analysis across two datasets and reports the delta.

UNII Band	Channels
UNII-1	36–48
UNII-2A/2C	52–96
UNII-3 (DFS)	100–148
UNII-4	149–177

---

Changelog

v1 → v2

Bug fixes

• Double-write bug fixed — v1 wrote every region file twice due to a logic error in the write loop, with the second write silently overwriting the first.

New features

• 46 countries instead of 4 coarse regions
• All 50 US states + DC with individual bounding boxes
• Channel analysis output per region and state
• Structured subdirectory output instead of flat folder
• --no-channel-summaries flag
• Improved terminal summary with top 15 countries and full state breakdown

v2 → v3

Bug fixes

• Small state fix — v3 introduces a bbox_area() helper that computes the area of each bounding box, then pre-sorts both COUNTRIES and US_STATES by area smallest-first at module load time. Classification then iterates the sorted lists so small states like Rhode Island match before larger overlapping neighbors like Massachusetts and Connecticut.

wigle_channel_compare.py (new)

Standalone comparison tool for A/B testing scan configurations. Accepts local device logs directly since WiGLE server exports strip channel data. See usage above.

---

Notes

• Offshore/ocean coordinates (e.g. Atlantic coast GPS drift, ferry captures) correctly land in other — this is expected behavior
• Coastal state bounding boxes (FL, NJ, NY, NC, SC, ME, MI) are extended slightly past land borders to capture barrier island and waterfront GPS
• wigle_no_gps.csv contains records with null or missing coordinates — excluded from all region files but preserved in the master deduped file
• WiGLE server-side CSV exports log -1 for Channel on all records — use local device exports for any channel analysis