The useful part of a Dune chart is the calculation you can inspect behind it. We followed a public DEX dashboard to its query and assessed access, data age and the measurements most easily misread.
The most useful feature in our Dune Analytics review was the route behind the chart. Clicking the volume ranking opened the calculation, its source table and its last execution details. Discovering that two time labels described different events taught us more than the polished chart itself.
What did we test on Dune?
On September 25, 2026, the KriptoMeta editorial team inspected the DEX metrics dashboard by @dune and volume query 4319 in a browser. We used the public pages without an account, opened the data-source menu and examined the query’s execution details.
We did not execute a new query, purchase a plan or connect a wallet. The assessment covers how well a reader can inspect existing results. It does not measure paid query performance or test the platform’s security infrastructure.
From a chart to its data source
The dashboard’s Data button lists tables used by its charts. We found dex.trades, dex_aggregator.trades and price tables in the menu. Visible source names mean readers do not have to rely solely on the dashboard creator’s description.
Dune / @dune DEX metrics, September 25, 2026. The Data menu on the right exposes source tables. Numbers are saved results visible at capture time.
We then opened the heading of the volume ranking. The query summed dollar values by DEX project, using dex.trades as its source and amount_usd in the calculation. Daily and weekly totals used separate time conditions.
- Identify the dashboard creator and the chart’s unit of measurement.
- Open the query through the chart heading.
- Compare its source table, date condition and counted field.
- Check when the result was produced and keep your interpretation within that period.
Access to the query is a substantial strength. Someone unfamiliar with SQL can still mistake finding the table for validating the data. Transparency makes an audit possible; it does not perform the audit for you.
A query date is not a data timestamp
The query we opened displayed Updated 1y for the edit history and Last run 2d ago for execution. Code can remain unchanged while results are refreshed. Conversely, a 24-hour condition in the code does not prove that a cached result covers the 24 hours ending now.
Dune query 4319. 1: last execution; 2: source table and time condition; 3: data source. Captured September 25, 2026, with markers added by KriptoMeta.
A NOW() condition is evaluated when the query runs. Viewing a saved result later does not move that window forward. Dune’s raw and processed data layers may also update at different times, so a recently executed query can still read a delayed table. The official freshness documentation treats those layers separately.
| Check | What it tells you | What it does not prove |
|---|---|---|
| Updated | When the content or query was edited | That the result was calculated now |
| Last run | When the query last executed | That the source contains the latest block |
| Date filter | The period requested by the calculation | That every dashboard card uses the same scope |
Where volume and user counts can mislead
Dune’s DEX data can record each pool-to-pool step separately. Consider a hypothetical 100 USDC swap routed through ETH before reaching the destination token. It can produce two pool records. If each leg is worth about $100, summing the rows gives roughly $200, although the user started with 100 USDC. We exclude fees and price changes only to keep the illustration simple.
The dex.trades documentation explains these individual steps. An aggregator table can represent the routed trade at a higher level. Since the same activity can appear in both tables, adding their volumes directly can double-count it. The appropriate measurement depends on whether you are studying pool activity or the user’s intended swap.
Address counts need similar care. One person may use several wallets, while automated activity can involve many addresses. A rising address count is not automatically a rising number of people or evidence of lasting adoption. Trading volume is not net capital inflow, protocol revenue or token-holder income either. The revenue and rights checks in our tokenomics guide help separate those questions.
From a dashboard to a finding worth citing
Before using a chart in reporting or analysis, keep more than the dashboard link. Record the query ID, measured field, chain filter, start and end times, and execution time. A claim about rising Ethereum DEX volume cannot rest on a query combining every chain. Comparing an incomplete current day with a complete previous day is another mismatch.
In a hypothetical check, dashboard A might show $100 while B shows $200. If A measures one user trade and B measures its two pool legs, the figures need not contradict each other. If both use the same table, period and filters, investigate execution time, additional joins and repeated rows instead.
Joins deserve particular attention: matching one trade row to three label rows and then summing the amount can count the same activity three times. A transaction hash is not a unique row key in every dataset: one transaction can contain multiple events. A chart that never explains its unit of observation remains weak evidence even when its total looks plausible.
A publication note could specify “Ethereum, DEX pool activity, September 1–7 UTC, query 4319,” with actual dates taken from the query being used. The dates here are illustrative, not a claim that we executed the reviewed query for that period. Preserve the query author and method explanation alongside the result.
Who is Dune suited to?
Readers looking for existing charts: public dashboards offer a quick route into a topic. The drawback is that different creators can use the same label for different measurements. Comparing two charts without matching the period, chains and units can produce a misleading conclusion.
Researchers who want to inspect the method: query and table links are a strength. If you need to investigate one transaction, a block explorer is usually the more direct starting point. Dune becomes useful when many records need to be grouped under a specific research question.
Readers comparing protocols: the ready-made TVL and fee screens covered in our DeFiLlama review can provide a quicker starting point. Dune offers more flexibility when you want to inspect the calculation or frame a different question. That flexibility also puts more responsibility on the reader to understand the method.
Free access and the limits of this review
At the time of our check, Dune’s pricing page described Free as a viewing tier. Query execution and scheduling were listed under Analyst, with private queries and dashboards under Plus. Reading a free chart does not imply unlimited computation or access to a private workspace.
For us, Dune’s value lies in opening a path from a result card to its source and method. A reader who treats a ready-made figure as an investment signal without examining it may miss that advantage. These screens document public access on a specific date; they do not demonstrate custom query performance, paid API usage or private workspace features.
