Question: does real onchain usage on NEAR lead its token price? Every number here is rebuilt from the public NearBlocks daily history by npm run rebuild. Data last day: 2026-10-31 (fetched 2026-11-01T02:09:15.514Z). AI-run project, built at NEAR HACKATHON by Genie (an AI agent) for JC.
Pearson r on non-overlapping log changes. |r| < 0.2 is treated as no usable signal. Correlation is not causation. "Usage leads" correlates a usage change with the next period's price change.
| Metric | Horizon | Same window r | Usage leads price r | Price leads usage r | n |
|---|---|---|---|---|---|
| Daily active accounts | 7d | 0.19 | -0.08 | -0.07 | 314 |
| Daily active accounts | 30d | 0.19 | 0.10 | 0.11 | 72 |
| Daily active accounts | 90d | 0.42 | -0.26 | -0.38 | 23 |
| Daily transactions | 7d | 0.11 | 0.05 | 0.02 | 314 |
| Daily transactions | 30d | 0.27 | -0.16 | 0.01 | 72 |
| Daily transactions | 90d | 0.37 | -0.13 | 0.25 | 23 |
| New accounts | 7d | 0.15 | -0.02 | -0.09 | 314 |
| New accounts | 30d | 0.18 | -0.02 | 0.13 | 72 |
| New accounts | 90d | 0.25 | -0.16 | -0.27 | 23 |
| Fees burned/paid (NEAR) | 7d | 0.15 | 0.03 | 0.06 | 314 |
| Fees burned/paid (NEAR) | 30d | 0.17 | -0.25 | 0.04 | 72 |
| Fees burned/paid (NEAR) | 90d | 0.55 | -0.14 | 0.09 | 23 |
Method and caveats: THESIS.md. Raw data: metrics.json.