Data in action / A field guide

See what the data can tell you.

Start with a question. Compare the evidence. Understand the limits. Explore three ways to turn research data into something you can actually use.

01

One asset. Different time horizons.

Is a short-window rise consistent with the longer window, or are they telling different stories?

Illustrative examples · not live market data

Change over each lookback window (%)

Five illustrative assets. Each column ends at the same snapshot, but looks back a different amount of time.

Asset changes by lookback window
Asset30m1h4h6h8h24h7d
BTC+0.18%+0.32%+0.75%+1.08%+1.25%+2.40%-3.60%
ETH-0.12%+0.08%+0.45%+0.71%+0.90%+1.70%-5.20%
SOL+0.35%+0.62%+1.40%+1.85%+2.30%+4.80%+2.10%
SUI-0.28%-0.46%-0.80%-1.10%-1.40%-2.30%+6.40%
ADA0.00%-0.05%+0.10%+0.08%+0.12%0.00%Missing

Scroll horizontally to compare all seven windows.

PositiveNegativeZeroMissing, not zero
Read across a row to compare horizons; read down a column to compare assets. Dark-to-bright intensity uses absolute change bands below 1%, 1–3%, and 3% or more. These are lookback windows, not a chronological price series.

In this invented example, BTC is up 2.40% over 24 hours but down 3.60% over seven days. Those observations can both be true; neither proves a reversal.

A useful next step

Prepare a daily research note that separates short-window moves from longer-window context, with the values and snapshot time attached.

What this cannot tell you

Whether an asset is fairly valued, why its price moved, or what it will do next. Different windows are not independent confirmations.

02

A broad move, or a few standouts?

Hold the time window constant. Compare the assets we track, without mistaking a five-asset sample for the whole market.

Illustrative examples · not live market data

24h change (%)
Compare over
3 positive1 negative1 zero0 missing
  1. SOL+4.80%
  2. BTC+2.40%
  3. ETH+1.70%
  4. ADA0.00%
  5. SUI-2.30%
Bars share a zero-centered, symmetric scale for the selected window; the scale adjusts when the window changes. Counts use the exact sign of each available value, not producer trend buckets. Equal values are ties; a high rank can still be negative.

A useful next step

Choose which observations deserve further research and explain that choice. A research assistant can use the same structured values behind this view.

Keep the universe visible

BTC, ETH, SOL, SUI and ADA are a bounded comparison set. Missing assets stay unknown. Volume-to-market-cap turnover is not order-book liquidity, and a high–low range is not realized volatility.

Visualize a response you already have

Paste a supported Asset Signal or Workbench market-regime response to replace the two market examples above. Processing happens only in this page's memory: no upload, browser storage, background refresh or purchase. The CFTC illustration below stays synthetic.

Use only data you are entitled to view. Paste the data response, never a private key, seed phrase, API credential or payment authorization. Maximum JSON size: 524,288 bytes. This checks supported structure, not source authenticity or payment.

Illustrative mode. Nothing has been imported.

03

Different participants. Different positions.

What does a weekly futures report show about the positions held by different participant categories?

Illustrative examples · not live market data

Net position / open interest (%)
Asset managersLeveraged funds

BTC

Asset managers+24%
Leveraged funds-18%

ETH

Asset managers+16%
Leveraged funds-9%

SOL

Asset managers+8%
Leveraged funds+3%
For illustration, net position is long minus short, divided by open interest. Left of zero is net short; right is net long. Compare categories at the same report date, not as if they were live spot-market flows. No actual report date is attached to these invented values.

A useful next step

Frame a question for a weekly positioning note: how do the participant groups differ, and how unusual is each position relative to its own available history?

Position is not intention

Net shorts may involve hedging or arbitrage. These reports do not cover all institutions, establish spot direction, or forecast a reversal. Historical percentiles also need their available sample size.

Read the regulatory positioning product guide

04

10 products. Four shared data themes.

Products package evidence for different jobs. They are not 10 independent sources, and several views reuse the same underlying observations.

Market observations

Asset Signal, Dirty Work, Agent Effort and Workbench.

Futures positions

Official CFTC report-derived positioning.

Company identity

Available Taiwan registration and factual company evidence.

Government procurement

Notices, awards and observed supplier–agency relationships.

Asset Signal

Compare one asset across time windows, then write a source-backed research note.

Not a valuation model or a buy/sell signal.

Regulatory Positioning

Compare reported futures positions across participant categories and report weeks.

Weekly futures evidence, not complete institutional flows.

Dirty Work Pack

Inspect identity, source lineage, freshness and missing evidence before quoting a number.

Data preparation and traceability, not an independent market source.

Agent Effort Metrics

Identify which inputs need review before an automated research workflow uses them.

Rule-based readiness; no measured time-saving claim.

Workbench Pack

Compare the five tracked assets and organize a research watchlist.

Shares underlying observations with the other market products.

Company Brief

Use a Taiwan business ID to establish company identity and available factual background.

Covered records only; missing data is not a clearance.

Supplier Lead Pack

Explore procurement evidence to find suppliers associated with a topic.

Candidates for further research, not supplier endorsements.

Tender Search

Find relevant notices and awards by topic, agency, supplier or case.

Distinguish historical awards from opportunities still open.

Company Activity

Review the government procurement cases linked to a known company.

Observed award amounts are not company revenue.

Supplier Pack

Examine observed agency concentration, annual activity and repeated award relationships.

Repeated awards do not by themselves prove renewals.

From a chart to a useful task.

A chart earns its place when it answers a real question, exposes its evidence, and makes the next research step clearer. Compare that workflow with your existing tools. A compelling picture alone is not proof of accuracy, investment performance, or willingness to pay.

Explore the crypto research products · Explore procurement applications