Skip to content

TensorStats — Comprehensive Metrics & Visualization Brainstorm

Date: 2026-02-14
Author: Hari subagent (metrics-brainstorm)

Legend

  • Complexity: Low / Med / High
  • Data Readiness:
  • Already ingested = in current DB tables/endpoints
  • Needs new pipeline = can be derived, but needs historical snapshots/materialized views/enrichment jobs
  • Needs chain RPC = requires additional runtime queries/event decoding beyond current ingest
  • Impact: Low / Med / High

Charting/UI stack suggestions: - Recharts: fast card charts, line/area/bar, lightweight tables + sparklines - Nivo: heatmaps, treemaps, choropleths, radar, swarm/bump - D3: custom alluvial, sankey, ridge plots, advanced brushing/interactions - Cytoscape.js: topology/network graphs - MapLibre/Deck.gl: geo clustering, arc/trail layers - Tremor: KPI cards, sparkline-heavy dashboards

0) Canonical Metric Inventory (Enumerate Everything Available)

0.1 network_state (per block snapshot)

  • total_subnets
  • active_subnets
  • total_nodes
  • active_nodes
  • total_stake
  • total_delegate_stake
  • total_issuance
  • total_overwatch_nodes
  • total_overwatch_stake

0.2 subnet_state (per subnet snapshot)

  • name
  • owner
  • total_stake
  • delegate_stake_balance
  • delegate_stake_shares
  • node_count
  • active_node_count
  • reputation
  • min_stake
  • max_stake
  • delegate_rewards_pct

0.3 node_state (per subnet × node snapshot)

  • hotkey
  • coldkey
  • peer_id
  • peer_multiaddr
  • bootnode_multiaddr
  • client_multiaddr
  • classification
  • classification_epoch
  • reputation
  • delegate_reward_rate

0.4 staker_state (per account × subnet snapshot)

  • stake amount
  • delegate_shares

0.5 hotkey_mapping

  • hotkey -> coldkey

0.6 network_peer

  • peer_id
  • roles
  • best_hash
  • best_number
  • first_seen
  • last_seen

0.7 node_ip_history

  • peer_id
  • ip
  • lat/lng
  • city
  • country
  • continent
  • observation_count
  • (plus first/last seen fields in canonical docs)

0.8 peer_geo

  • peer_id
  • ip
  • lat/lng
  • country
  • region
  • city
  • isp
  • asn
  • timezone

0.9 subnet_bootnode

  • subnet_id
  • multiaddr
  • peer_id
  • source

0.10 peer_multiaddr

  • peer_id
  • multiaddr
  • subnet_id
  • node_id
  • source

0.11 PoI / Attestation (validator_epoch-derived)

  • total attestations per validator per epoch
  • passed per validator per epoch
  • failed per validator per epoch
  • per-subnet consensus participation rates

0.12 ETH DuckDB side

  • top holders
  • transfers
  • wallet flow graph
  • wallet timeline with buy/sell markers
  • genesis cohort
  • lot-based FIFO P&L

1) network_state Brainstorm

Derived/computed metrics

  • Subnet activity ratio = active_subnets / total_subnets
  • Node activity ratio = active_nodes / total_nodes
  • Delegation penetration = total_delegate_stake / total_stake
  • Overwatch penetration (nodes) = total_overwatch_nodes / total_nodes
  • Overwatch stake share = total_overwatch_stake / total_stake
  • Stake-per-active-node, stake-per-subnet
  • Issuance velocity (1d/7d rolling delta)
  • Issuance acceleration (second derivative)
  • Regime shift flags (CUSUM/change-point on stake growth)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Network health ribbon (5 mini sparklines: active ratio, delegation share, issuance growth, overwatch share, stake/node) Sparkline strip + KPI cards Low Already ingested High Home Tremor/Recharts
Active vs total subnets/nodes over time Dual-line + filled gap area Low Already ingested High Home Recharts
Stake composition (self vs delegated vs overwatch) Stacked area Low Already ingested High Home Recharts
Issuance momentum dashboard Line + histogram of block deltas Med Needs new pipeline (windowed aggregates) High Home/Token Nivo/Recharts
Network stress gauge (composite z-score) Gauge/donut + trend sparkline Med Needs new pipeline Med Home Tremor
Contribution waterfall (which component changed network health over 24h) Waterfall Med Needs new pipeline Med Home D3

Cross-metric correlations

  • Node activity ratio vs global PoI pass rate (lagged correlation)
  • Delegation penetration vs subnet reputation median
  • Overwatch stake share vs attestation failure rate
  • Issuance velocity vs new subnet activation rate

Time-series opportunities (track historically)

  • Daily/epoch snapshots of all ratios (not just totals)
  • Rolling volatility of node activity ratio
  • Trend and seasonality decomposition for issuance
  • Alert baselines (95% bands) for sudden activity drops

2) subnet_state Brainstorm

Derived/computed metrics

  • Subnet active node ratio = active_node_count / node_count
  • Delegation share = delegate_stake_balance / total_stake
  • Stake concentration across subnets (HHI, Gini)
  • Reputation efficiency = reputation / log(total_stake)
  • Subnet capacity spread = max_stake - min_stake
  • Stake accessibility index = min_stake / median_stake (across subnets)
  • Reward attractiveness score (delegate_rewards_pct adjusted by reputation and activity)
  • Subnet growth rate (stake, nodes, reputation deltas)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Stake market map of subnets Treemap (size=stake, color=reputation) Low Already ingested High Subnets Nivo
Risk/return scatter Bubble scatter (delegate_rewards_pct vs reputation, size=stake) Low Already ingested High Subnets Recharts/Nivo
Subnet leaderboard table with inline sparklines Table + sparklines + delta badges Med Needs new pipeline (history) High Subnets/Home Tremor
Subnet health radar Radar (activity, rep, stake growth, delegation share, decentralization proxy) Med Needs new pipeline Med Subnet detail Nivo
Rank race Bump chart for top-N subnets by stake/reputation over time Med Needs new pipeline High Subnets Nivo
Distribution quality Box/violin of reputation by subnet cohorts (small/medium/large stake) Med Needs new pipeline Med Subnets/Research view Nivo/D3
Owner concentration view Sankey owner → subnet → stake High Already ingested Med Subnets D3

Cross-metric correlations

  • total_stake vs reputation
  • active_node_count vs reputation
  • delegate_rewards_pct vs delegation inflows
  • min_stake/max_stake spread vs subnet growth/churn
  • subnet reputation vs validator PoI performance within subnet

Time-series opportunities

  • Subnet lifecycle states: launched → growth → plateau → decline
  • Delegation inflow/outflow per subnet over epochs
  • Reputation velocity and drawdown statistics
  • Persistent subnet “quality score” history for ranking stability

3) node_state Brainstorm

Derived/computed metrics

  • Node active flag by classification mapping
  • Classification tenure = current_epoch - classification_epoch
  • Reputation velocity and volatility per node
  • Delegate reward competitiveness percentile (within subnet)
  • Endpoint redundancy score (count of populated multiaddr fields)
  • Node stability score (classification switches + reputation variance)
  • Hotkey reuse rate / coldkey fanout (join hotkey_mapping)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Node class transition flow Alluvial/Sankey across epochs (Idle→Validator etc.) High Needs new pipeline (historical states) High Validators/Subnets D3
Reputation by class Box/violin grouped by classification Low Already ingested High Validators Nivo
Reward rate vs reputation efficiency Scatter + regression band Low Already ingested High Validators Recharts
Per-node timeline strip Gantt-like classification_epoch track Med Needs new pipeline Med Validator detail D3
Reputation movers Lollipop (top gainers/losers over 7d) Med Needs new pipeline High Validators/Home Recharts
Endpoint health donuts Donuts for % nodes missing peer/bootnode/client multiaddr Low Already ingested Med Topology/Validators Tremor
Subnet-node network with class coloring Force/concentric network Med Already ingested Med Topology Cytoscape.js

Cross-metric correlations

  • Delegate reward rate vs node reputation
  • Classification tenure vs pass/fail outcomes (PoI join)
  • Endpoint redundancy vs network_peer lag
  • Node reputation vs geo/ASN concentration exposure

Time-series opportunities

  • Store node reputation per epoch/block
  • Track classification transitions and dwell time
  • Longitudinal node reliability score
  • Survival curves for new nodes becoming validators

4) staker_state Brainstorm

Derived/computed metrics

  • Delegator count per subnet
  • Average/median stake size per subnet
  • Stake concentration by subnet and globally (HHI, Gini, Theil)
  • Whale share (top 1/5/10 delegators)
  • Net delegation flow (in/out by wallet/subnet over time)
  • Delegate share efficiency (delegate_shares / stake normalization)
  • Delegator retention/churn cohorts

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Wealth concentration lens Lorenz curve + Gini card Med Already ingested High Token/Subnets D3/Recharts
Delegation routing Sankey staker cohort → subnet → class tier High Needs new pipeline (cohorting) High Subnets/Token D3
Stake-size distribution Histogram + log-scale Low Already ingested Med Token/Validators Recharts
Staker-subnet exposure matrix Heatmap (top wallets x subnets) Med Already ingested High Token wallets Nivo
Cohort survival Retention heatmap by first-seen epoch High Needs new pipeline High Token wallets D3/Nivo
Whale movement monitor Streamgraph of top-N wallet net flow High Needs new pipeline High Token/Home D3

Cross-metric correlations

  • Delegator concentration vs subnet reputation volatility
  • Whale concentration vs consensus participation stability
  • Delegate share efficiency vs validator reward rates
  • Net flow vs changes in delegate_rewards_pct

Time-series opportunities

  • Daily stake balance snapshots by wallet×subnet
  • Cohort entry/exit trends
  • Flow decomposition: new stake vs migrated stake vs withdrawn stake

5) hotkey_mapping Brainstorm

Derived/computed metrics

  • Hotkeys per coldkey
  • Coldkey concentration index (entity control centralization)
  • Cross-subnet operator footprint (if hotkeys span subnets)
  • Sybil-risk heuristic (many hotkeys with synchronized behavior)
  • Ownership churn rate (mapping changes over time)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Operator footprint panel Bar chart: hotkeys per coldkey (Top 20) Low Already ingested High Validators Recharts
Ownership topology Bipartite graph coldkey ↔ hotkey Med Already ingested Med Topology/Validators Cytoscape.js
Concentration dial HHI/Gini gauge for ownership centralization Med Needs new pipeline (time history) Med Home/Validators Tremor
Cross-subnet control map Sunburst coldkey → subnet → node count High Already ingested (join with node_state) High Validators/Subnets Nivo
Potential sybil cluster explorer Similarity graph (behavioral clustering) High Needs new pipeline Med Research/Validators D3

Cross-metric correlations

  • Coldkey centralization vs PoI failure clustering
  • Operator footprint vs average node reputation
  • Shared coldkey operators vs geographic concentration

Time-series opportunities

  • Mapping change events timeline
  • Operator expansion/contraction trend
  • Coldkey-level risk score trend

6) network_peer Brainstorm

Derived/computed metrics

  • Peer recency score (now - last_seen)
  • Online duration (last_seen - first_seen)
  • Head lag = network_tip - best_number
  • Stale peer ratio (lag > threshold)
  • Hash divergence ratio (fraction not on modal best_hash)
  • Role distribution entropy

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Head lag distribution Histogram + percentile bands Low Already ingested High Topology/Home Recharts
Peer health monitor Status grid (online/stale/divergent) Med Already ingested High Topology Tremor
Hash consensus snapshot Treemap by best_hash cluster sizes Med Already ingested Med Topology Nivo
Roles over time Stacked area of role counts Med Needs new pipeline (historical snapshots) Med Topology Recharts
Peer session raster Heatmap timeline (peer x time, seen/not seen) High Needs new pipeline High Topology/Validators D3/Nivo
Outlier watchlist Ranked table with lag sparkline per peer Med Needs new pipeline High Topology Tremor

Cross-metric correlations

  • Head lag vs attestation fail rates
  • Peer role vs reputation/classification
  • Divergent hash clusters vs geo/ASN segments

Time-series opportunities

  • Rolling peer availability SLO (e.g., 99% seen)
  • Session-length distributions over time
  • Episode detection for network partitions

7) node_ip_history Brainstorm

Derived/computed metrics

  • IP churn rate per peer
  • Geo churn rate (city/country transitions)
  • Mobility distance traveled (great-circle cumulative)
  • Observation-weighted presence by region
  • Residential vs hosting proxy heuristic (with ASN enrichment)
  • Stability tier labels (static / occasional / highly volatile)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Geographic footprint Choropleth + bubble overlay by observation_count Med Already ingested High Topology/Home MapLibre/Nivo
Node movement trails Arc/trail map per peer over time High Needs new pipeline (ordered history API) High Topology Deck.gl
IP churn leaderboard Bar + sparkline (changes per week) Med Needs new pipeline High Validators/Topology Recharts
Country transition flow Sankey source-country → destination-country High Needs new pipeline Med Topology D3
Stability heatmap Peer x week heatmap of IP changes High Needs new pipeline Med Topology Nivo

Cross-metric correlations

  • IP churn vs PoI pass/fail volatility
  • Region mobility vs classification instability
  • Country concentration vs network_peer hash divergence

Time-series opportunities

  • Persist ordered IP observations with timestamps in API
  • Weekly geo concentration trend
  • Detect “sudden migration events” (many peers relocating)

8) peer_geo Brainstorm

Derived/computed metrics

  • Country/region/city share distribution
  • ASN concentration (HHI/Gini)
  • ISP concentration and systemic exposure
  • Timezone dispersion index
  • Geo decentralization score (entropy-based)
  • Jurisdiction risk flags (custom risk map)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Decentralization globe 3D/2D map with cluster density Med Already ingested High Topology Deck.gl/MapLibre
ASN dependency map Treemap (size=peer count, color=risk/lag) Med Already ingested High Topology/Validators Nivo
Timezone wheel Radial bar chart by timezone counts Low Already ingested Med Topology Nivo
Geo diversity scorecards KPI cards + historical sparkline Med Needs new pipeline (historical snapshots) High Home/Topology Tremor
Country-ISP bipartite Network graph to show concentration chokepoints High Already ingested Med Topology Cytoscape/D3

Cross-metric correlations

  • ASN concentration vs network lag or consensus failures
  • Geo decentralization score vs subnet reputation stability
  • Timezone coverage vs epoch participation consistency

Time-series opportunities

  • Historical geo snapshots (daily)
  • ASN dominance trend
  • Regional failover behavior during incidents

9) subnet_bootnode Brainstorm

Derived/computed metrics

  • Bootnode count per subnet
  • Unique peer_id count per subnet
  • Bootnode redundancy index
  • Bootnode geo/ASN diversity (join peer_geo)
  • Bootstrap centralization risk (single source/single ASN)
  • Source reliability score (source stability over time)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Bootnode resilience panel Gauge + counts (redundancy, diversity) Med Already ingested (with joins) High Subnet detail/Topology Tremor
Subnet↔bootnode map Bipartite graph Med Already ingested High Topology Cytoscape
Bootnode source mix Stacked bar by source per subnet Low Already ingested Med Subnets Recharts
Bootstrap risk heatmap Heatmap subnet x risk factors Med Needs new pipeline (scoring) High Subnets/Topology Nivo
Bootnode drift timeline Event timeline when bootnodes change High Needs new pipeline Med Subnet detail D3

Cross-metric correlations

  • Bootnode redundancy vs active_node_count
  • Bootnode diversity vs peer head lag
  • Source concentration vs subnet PoI participation variance

Time-series opportunities

  • Bootnode set change history
  • Mean time between bootnode changes
  • Incident overlays (failures after topology contractions)

10) peer_multiaddr Brainstorm

Derived/computed metrics

  • Multiaddr count per peer
  • Protocol composition (tcp/ws/quic/etc parsed from multiaddr)
  • Subnet overlap score (peer appearing across subnets)
  • Address entropy (diversity/uniqueness)
  • Source consistency score
  • Reachability proxy (address completeness heuristics)

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Protocol mix dashboard Stacked bars by protocol and subnet Med Needs new pipeline (protocol parsing) Med Topology Recharts
Source→protocol→subnet flow Sankey High Needs new pipeline Med Topology D3
Address richness by peer Dot plot / beeswarm Med Already ingested Med Validators/Topology Nivo
Multiaddr quality badge in tables Sparkline + categorical badge Low Already ingested High Validators/Subnets Tremor
Peer-subnet overlap chord Chord diagram for multi-subnet peers High Needs new pipeline Med Topology D3

Cross-metric correlations

  • Address richness vs node reputation
  • Protocol diversity vs peer lag/stability
  • Subnet overlap vs operational centralization risk

Time-series opportunities

  • Address additions/removals per peer
  • Protocol adoption trend over time
  • Stability of source attribution

11) PoI / Attestation Brainstorm (Priority Section)

Derived/computed metrics

  • Pass rate = passed / total
  • Failure rate = failed / total
  • Miss rate (if total implies expected, derive missing submissions) (needs RPC clarity)
  • Epoch z-score vs validator’s trailing mean
  • EWMA pass-rate trend
  • Pass/fail streak lengths (max, current)
  • Drawdown from personal best pass rate
  • Subnet consensus quality index:
  • weighted(pass_rate, participation_rate, volatility_penalty)
  • Validator reliability tier (A/B/C/D)
  • Surprise failure score (actual fails - expected fails)
  • Time-to-recovery after a failure cluster
  • Timeliness buckets (on-time/late/missed) (needs attestation timestamp pipeline)

Core visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Validator × Epoch pass/fail heatmap Heatmap (green/red intensity; tooltip counts) Med Already ingested (epoch-level data) High Validators Nivo
Streak analysis panel Barcode strip + streak length distribution Med Needs new pipeline (derived streak table) High Validators/Validator detail D3
Consensus quality scoring leaderboard Ranked table + sparkline + badge Med Needs new pipeline High Validators/Home Tremor/Recharts
Attestation timeliness distribution Histogram/violin of latency buckets High Needs chain RPC + new pipeline High Validators Nivo/D3
Subnet participation matrix Heatmap subnet × epoch participation Med Already ingested High Subnets/Validators Nivo
Failure cascade timeline Event timeline with cluster detection High Needs new pipeline High Validators/Home D3
Reliability control charts Run chart with control limits for top validators Med Needs new pipeline Med Validator detail Recharts
Rank momentum race Bump chart validator quality rank over time Med Needs new pipeline High Validators Nivo
Distribution drift Ridgeline by epoch of validator pass rates High Needs new pipeline Med Validators/Research D3
Ternary reliability plot Ternary (pass/fail/missed) High Needs chain RPC for misses/timeliness Med Validators D3

PoI-specific cross-metric correlations

  • Pass rate vs validator reputation
  • Pass rate vs delegated stake share
  • Failure volatility vs node classification churn
  • Subnet participation rate vs subnet active node ratio
  • Timeliness (when available) vs peer head lag (network_peer.best_number lag)
  • Consensus quality vs geo concentration (country/ASN)

PoI time-series opportunities

  • Persist epoch-level per-validator panel (long horizon)
  • Rolling reliability percentiles (7d/30d)
  • Regime change alerts (sudden consensus degradation)
  • Incident replay mode (align failures with topology/geo changes)

12) ETH DuckDB Analytics Brainstorm

Derived/computed metrics

  • Holder concentration (HHI/Gini, top-N share)
  • Net transfer flow by wallet cohort
  • Whale inflow/outflow momentum
  • Wallet graph centrality (in/out degree, PageRank)
  • Buy/sell win rate and holding period stats
  • FIFO realized P&L, unrealized P&L, tax-lot aging
  • Genesis cohort retention and behavior drift
  • Profit factor by wallet segment

Visualization ideas

Idea Viz type Complexity Data readiness Impact Page Library
Holder concentration cockpit Lorenz curve + treemap + top-N share cards Med Already ingested High Token/Holders D3/Nivo
Transfer river Sankey/alluvial by wallet cohorts (exchange, whale, retail) High Needs new pipeline (entity labels/cohorts) High Token D3
Wallet flow graph explorer Force graph with filtering + edge bundling Med Already ingested High Token/Wallets Cytoscape/D3
Wallet timeline trading tape Candlestick-style buy/sell markers + inventory line Med Already ingested High Wallet detail Recharts
P&L decomposition Waterfall (realized gains/losses, fees, unrealized) Med Already ingested High Wallet detail Recharts
Cohort retention matrix Heatmap genesis cohort retention by month Med Already ingested High Token Nivo
Lot age risk chart Histogram/box of lot age vs unrealized return Med Already ingested Med Wallet detail Nivo
Smart-money tracker Scatter (wallet return vs turnover; size=AUM proxy) High Needs new pipeline (segment heuristics) Med Token/Wallets Recharts

Cross-metric correlations

  • Holder concentration vs transfer volatility
  • Whale netflow vs price/issuance proxies (if market data integrated)
  • Wallet centrality vs profitability
  • Genesis cohort behavior vs broader token distribution shift
  • ETH wallet inflows vs Substrate delegation growth (cross-chain narrative)

Time-series opportunities

  • Daily holder distribution snapshots
  • Rolling netflow by cohort and cluster
  • Longitudinal wallet factor model (momentum, mean-reversion, accumulation)

13) Cross-Domain “Quant Desk” Views (Multi-table joins)

13.1 Composite scores

  • Network Resilience Index = weighted(geo diversity, peer lag health, bootnode redundancy, node activity)
  • Complexity: Med
  • Readiness: Needs new pipeline (composite model)
  • Impact: High

  • Page: Home

  • Subnet Investability Score = weighted(reputation, pass-rate stability, delegation depth, churn penalty)

  • Complexity: High
  • Readiness: Needs new pipeline
  • Impact: High

  • Page: Subnets

  • Validator Quality Score = weighted(attestation success, timeliness, reputation trend, uptime proxy)

  • Complexity: High
  • Readiness: Needs chain RPC + new pipeline
  • Impact: High
  • Page: Validators

13.2 Correlation lab (interactive)

  • Scatter matrix (pair plot) with selectable x/y/color/size metrics across network/subnet/node/PoI
  • Complexity: High
  • Readiness: Needs new pipeline (feature table)
  • Impact: High
  • Page: New “Research” tab or advanced mode in Validators/Subnets
  • Library: D3 + crossfilter

13.3 Regime & anomaly detection

  • Change-point overlays on stake, pass-rate, and peer lag
  • Residual alerts (actual vs expected consensus quality)
  • Complexity: High
  • Readiness: Needs new pipeline
  • Impact: High
  • Page: Home + Validators
  • Tier 1 (Home): simple KPI cards + sparklines + alert chips
  • Tier 2 (List pages): sortable tables with inline microcharts (sparkline, mini bar, tiny heat cells)
  • Tier 3 (Detail pages): deep distributions, cohort analysis, event timelines
  • Tier 4 (Research mode): custom metric builder + correlation lab + downloadable datasets

Useful components: - “Metric Pills” to toggle denominator normalization (absolute vs per-node vs per-stake) - Timebrush for shared cross-chart zoom - Small-multiple grids for easy subnet/validator comparison - Drill-down flow: Home anomaly → subnet outlier → validator timeline → peer geo evidence

15) Immediate Build Priorities (Highest ROI)

  1. PoI Validator×Epoch Heatmap + Streaks + Quality Leaderboard
  2. Complexity: Med
  3. Readiness: Mostly already ingested (streaks need derived table)
  4. Impact: Very high

  5. Page: Validators

  6. Subnet Risk/Return Scatter + Treemap + Sparkline Leaderboard

  7. Complexity: Med
  8. Readiness: Core data already ingested; history pipeline needed for sparklines
  9. Impact: Very high

  10. Page: Subnets

  11. Network Health Ribbon on Home

  12. Complexity: Low
  13. Readiness: Already ingested
  14. Impact: High

  15. Page: Home

  16. Geo/ASN Decentralization Panel

  17. Complexity: Med
  18. Readiness: Already ingested (point-in-time), history needed for trends
  19. Impact: High

  20. Page: Topology/Home

  21. ETH Holder Concentration + Wallet P&L Decomposition Enhancements

  22. Complexity: Med
  23. Readiness: Already ingested
  24. Impact: High
  25. Page: Token + Wallet detail

16) Data Engineering Backlog to Unlock Advanced Visuals

  • Create historical snapshot fact tables for subnet/node/geo/peer dimensions (daily + epoch granularities)
  • Build derived metric marts:
  • fact_validator_poi_epoch
  • fact_subnet_daily
  • fact_peer_health_hourly
  • fact_staker_flow_daily
  • Add parser jobs for multiaddr protocol extraction
  • Add attestation timeliness ingestion (requires extra chain fields/events)
  • Add anomaly/regime detector jobs for alert overlays
  • Materialize top-N pre-aggregates for snappy frontend filters

17) Notes on Library Fit

  • Use Recharts/Tremor for 80% dashboard primitives and fast shipping.
  • Use Nivo for heatmaps/treemaps/radar/bump with good theming speed.
  • Use D3 only for bespoke quant visuals (alluvial, ridgeline, ternary, advanced brushing).
  • Keep Cytoscape for topology graph semantics.
  • Use MapLibre/Deck.gl when geo density/trails become core.