From market data to a research decision people can review

A chart can be visually precise and still leave important questions unanswered. Which venue supplied the observations? What does the timestamp represent? Was an empty interval excluded or replaced? A research workflow becomes more dependable when those questions travel with the data, rather than remaining in one analyst's memory. The objective is not to eliminate uncertainty. It is to make the evidence, transformations and limitations clear enough for someone else to review.

Define the research question first

Start with a question that can be answered by the available material. “What changed in this venue's reported trading activity during the selected period?” is more precise than “What is the market doing?” The first question has a defined source and scope. The second may require information that a single feed cannot provide.

Set the observation window, comparison period and intended audience before collecting data. Record exclusions, such as instruments without sufficient history. This keeps a convenient dataset from silently deciding the question. It also makes it easier to distinguish an exploratory observation from a conclusion that has been checked across several sources.

Build a small source record

For each dataset, retain the provider, collection time, instrument identifier, venue, quotation currency and measurement units. State whether a time label represents an event, a publication or your own collection. If a figure is calculated, describe the inputs and method. Keep the original observation separate from the transformed version so a reviewer can follow the calculation.

These details matter even for familiar charts. Coinbase's Exchange candle documentation explains that historical data may be incomplete and that intervals without ticks are not published. That is a specific provider behavior, not a rule for every feed. It illustrates why an absent row should not automatically become a zero price or a confirmed outage.

Make quality checks visible

A compact review can check duplicated records, unexpected gaps, mismatched units and observations outside the selected window. If a figure differs from a second source, first compare definitions and collection times. A last-trade price, an indicative quote and a calculated reference value need not describe the same thing. Label the difference before assuming one source is wrong.

When an input fails a check, decide explicitly whether to exclude it, investigate it or retain it with a qualification. Record that choice near the output. Filling a gap may make a chart easier to draw, but it creates a derived value. Readers should be able to distinguish that value from an observed one. If the remaining evidence does not support the question, say so.

Preserve what was known at the time

A historical analysis can change when its sources are revised. This is especially relevant when market research incorporates economic data. The St. Louis Fed's FRED and ALFRED documentation distinguishes information available today from information known during a past period. Its vintage-date documentation identifies releases that introduced or revised observations.

Our practical recommendation is to record the version or retrieval date used in a note and retain enough context to reproduce the result. If a later revision changes the conclusion, update the note with a clear explanation. Quietly replacing the underlying figures leaves readers unable to tell whether the original analysis was mistaken or the source itself changed.

Turn the analysis into a reviewable note

A useful note can follow five short parts: the question, the observed evidence, the method, the interpretation and the limitations. Keep factual observations distinct from possible explanations. Include alternative explanations when the evidence cannot distinguish between them. State what additional information would strengthen or weaken the interpretation.

For example, a hypothetical rise in a venue's reported volume could justify further investigation; by itself it does not establish the cause, the quality of that activity or a future price direction. A concise note that acknowledges this boundary is more useful than a confident forecast unsupported by its inputs. This example is educational and does not recommend a trading decision.

Make the final handoff explicit

Before publication, confirm that the reviewer can open the sources, understand the selected period and trace at least one representative calculation. Assign a version and name the person responsible for corrections in the actual publishing process. Store unresolved questions with the note rather than burying them in a separate conversation.

The result is a research record that remains useful after the immediate discussion ends. Shared definitions, visible uncertainty and traceable changes help a team learn from its work. They do not make incomplete data complete, but they make the limits easier to recognize before an observation becomes a decision.