5 ms·

I designed and built an airport weather sensor network. Output is available here: https://davidsen.aero/awas/kgbr/ https://davidsen.aero/awas/kgbr/ (Still work

I designed and built an airport weather sensor network. Output is available here: https://davidsen.aero/awas/kgbr/ https://davidsen.aero/awas/kgbr/ (Still working on vis, sky condition, and present weather.) As you can see by some missing data on the graphs, it’s not always available. But the sensors are carefully selected to perform to FAA standards. It’s been really fun putting together a high-precision, distributed system (packet radio data links) that can withstand the outdoors.

chrisdalke 6y ago
Nice work! How are you persisting the data? Are you just keeping the history that the website shows, or keeping all historical data?
- danaliv 6y ago
  Thanks! The controller (box that runs the network and all the algos) keeps a day’s worth of raw sensor data, intermediate calculations, and final user-ready observations. The data behind the web interface is just observations and never gets pruned. In both cases, data is in SQLite.
  - chrisdalke 6y ago
    Sounds exciting! related, I've been building a open-source platform for hardware telemetry data called TelemetryJet (https://www.telemetryjet.com/ https://www.telemetryjet.com/). It provides a toolkit for data collection, analysis, and visualization. The aim is to make it easy to build a long-term time series database, share it with collaborators, and actually gain insight from the data you collect. I've been mostly focused on data from experimental electric vehicles, but I am also really interested in hearing your experience building a sensor network. I'd really love to hear what kinds of challenges you encountered building a distributed sensor system like this, particularly with collecting and sharing the sensor data. If you're willing, please reach out to me at the email in my bio!
    - danaliv 6y ago
      This looks fantastic, and is precisely the sort of thing I’ve wished I had for this system. As it is, the data collection is very much “MVP.” I’ll definitely get in touch!
      - chrisdalke 6y ago
        Realized I didn't actually have it in the profile and there's some sort of filtering on emails or the profile isn't updating. My email is my profile username @ gmail!

chrisdalke 6y ago

Nice work! How are you persisting the data? Are you just keeping the history that the website shows, or keeping all historical data?

danaliv 6y ago
Thanks! The controller (box that runs the network and all the algos) keeps a day’s worth of raw sensor data, intermediate calculations, and final user-ready observations. The data behind the web interface is just observations and never gets pruned. In both cases, data is in SQLite.
- chrisdalke 6y ago
  Sounds exciting! related, I've been building a open-source platform for hardware telemetry data called TelemetryJet (https://www.telemetryjet.com/ https://www.telemetryjet.com/). It provides a toolkit for data collection, analysis, and visualization. The aim is to make it easy to build a long-term time series database, share it with collaborators, and actually gain insight from the data you collect. I've been mostly focused on data from experimental electric vehicles, but I am also really interested in hearing your experience building a sensor network. I'd really love to hear what kinds of challenges you encountered building a distributed sensor system like this, particularly with collecting and sharing the sensor data. If you're willing, please reach out to me at the email in my bio!
  - danaliv 6y ago
    This looks fantastic, and is precisely the sort of thing I’ve wished I had for this system. As it is, the data collection is very much “MVP.” I’ll definitely get in touch!
    - chrisdalke 6y ago
      Realized I didn't actually have it in the profile and there's some sort of filtering on emails or the profile isn't updating. My email is my profile username @ gmail!