Datamining Meshtastic - Statistics of a small sized network in Southern Germany

 Date: January 10, 2025

From 31.12.2024 - 10.1.2024 I was running a heavily modified Meshlytics on my Meshtastic station.

The setup was rather simple: 5dBi omnidirectional antenna mounted onto a T-Beam that was connected into my local Wi-Fi to extract MQTT. Everything was placed onto my small balcony next to my office.

Simulated coverage map
Simulated coverage map

My location is in a deep valley - so only 3-4 stations with very low signal levels (-120dBm to -145dBm) are reachable. These stations are placed in strategically good positions by the local community, offering me a gateway.

SNR from directly received stations
SNR from directly received stations

Looking at the relation RSSI vs. distance to my station, it looks even worse. Not that dense point cloud as not every station is sending their GPS position.

Plot RSSI vs. distance from my home station
Plot RSSI vs. distance from my home station. Distance is in kilometer and RSSI in dBm.

But let’s have a look at the number of stations I can see from here.

Amount of stations received
Amount of stations received

And the types of data being put on the waves. You can see: our network is not really chatty. Position reports, Nodeinfos, and Telemetry are dominating the game.

Packet types being used in Meshtastic
Packet types being used in Meshtastic

The valley position has the downside that all stations are far away in terms of hops.

Received stations and their hops over time
Received stations and their hops over time. Negative values means that the sending station uses a broken firmware.

The RSSI plot of the entire network shows: it lives from the robustness of LoRa.

RSSI of all stations (all hops_away included). It shows the RSSI of the last station, that received/forwarded the packet  in my case these are the weak stations from the first picture.
RSSI of all stations (all hops_away included). It shows the RSSI of the last station, that received/forwarded the packet - in my case these are the weak stations from the first picture.

Last but not least, the channel utilization and TX time reports might be interesting for comparison.

Channel Utilization and TX time used
Channel Utilization and TX time used

Previous
⏪ MicRadar R24BBD1: Reverse Engineering another 24GHz CW radar

Next
Zephyr: Missing USART4 definition in .dtsi files of the STM32U0 ⏩