GCP2 Experiment Console - Open-Source Live Node Monitoring & Experiment Logging

I’ve now published the GCP2 Experiment Console on GitHub as an open-source project.

What began as a tool for monitoring my own GCP2 node 58114 has developed into a multi-node console that can be used with other GCP2 nodes as well.

GitHub:
https://github.com/contactproject/GCP2-Experiment-Console

The console records the selected node’s Device Coherence, the Global Network Variance, timestamped START / EVENT / END markers, experiment notes, and local history for later analysis. It also provides CSV/JSONL export and database backup/restore.

The purpose is not to interpret anomalous readings automatically, but to make experiments easier to predefine, timestamp, preserve and reproduce.

I’m particularly interested in whether other GCP2 participants can run it against their own nodes and report what works, what breaks, and what would make it more useful.

**Current release: v0.4.2
**

I got it running on my own node - 498.

Fantastic - this is exactly what I was hoping for. Really good to see the console actually running on another GCP2 node!

Thanks for giving it a try on node 498. If you run into anything odd, or have ideas for making it better, please let me know. The more people test it on their own nodes, the more useful this can become.

1 Like

A small but interesting result from our first marked GCP2 experiment with node 58114.

During a predefined 15-minute experimental period, the local node measure rose from an average of 31.38 to 51.03 — about a 63% increase. After the experiment ended, it fell back to 30.07, almost exactly where it had started.

That pattern caught our attention.

However, the wider GCP2 network also showed a strong increase during the same period, so we cannot say the change was caused by our local experiment.

In other words: something changed during the marked interval, but we do not yet know why.

GCP2 Experiment Console — v0.5.1 update

I’ve updated the console with formal trial declarations, a persistent trial register, improved exports and backups, and a Stop & close button.

The methodological changes follow feedback from Roger Nelson, who wrote:

“Pre-declaration and the transparency (keeping results from all formally specified tests) you mention are both essential to meet the challenges of low signal-to-noise experiments.”

Formal trials now require the hypothesis, duration, controls, analysis, and stopping rule before START. The declaration then locks; subsequent notes and outcome revisions are appended. Negative, inconclusive, and interrupted trials remain in the record.

Existing observations without complete declarations remain exploratory. Roger’s input concerned the methodology, rather than validation of the software or hypothesis.

Hide continues collection; Stop & close stops collection while preserving data. The console still records derived AC/netvar statistics, not raw RNG channels.

Source and documentation on GitHub

Back up your database before updating. Feedback and testing on other nodes are welcome!

The next version needs to include a countdown timer and optional parameters instead of mandatory. A hack is to not fill out the fields by inserting a dash.