Key idea
A data quality check can be more valuable than another indicator when the input itself is uncertain.
A number can be precise and still answer the wrong question. A quote for the wrong expiration, a percentage using a different baseline, or a missing value displayed as zero can all produce a neat-looking research mistake. Before interpreting data, establish what the number represents.
Use the following checks whenever something looks unusually attractive, contradictory, or too clean. The goal is a short inspection you can repeat, not a complicated audit that prevents every research session.
Confirm identity and time
Check the exact instrument. For an option, confirm underlying symbol, expiration, strike, and call or put. For an index or fund, confirm that you are comparing the intended product. Similar names do not mean identical exposure.
Next, check the timestamp and session. A closing value, an extended-hours observation, and a current regular-session quote should be labeled distinctly. If the display does not reveal the relevant time, record that limitation before drawing a time-sensitive conclusion.
Verify units and coverage
Ask whether a field represents shares, contracts, dollars, a percentage, or a calculated score. A hypothetical “500” might mean 500 contracts rather than $500 of premium. A percentage might describe change since the previous close rather than change since your first observation.
Also identify coverage. Does a summary include all expirations, selected contracts, one venue, or a filtered sample? A total can only describe the population included in the calculation. Comparing totals with different filters can create a false impression of agreement or disagreement.
Inspect adjustments and missing values
Historical price series may use different adjustment conventions. A corporate action can create a discontinuity that needs explanation before you label it an extraordinary move. When comparing sources, check whether their series are adjusted in the same way.
Distinguish zero from unavailable. Zero means a measured quantity was zero under the provider's definition. Unavailable means the observation is missing or not supplied. Neither should silently replace the other in a calculation.
Read the definition and cross-check the outlier
Finally, read the provider's description of derived fields. If a label says bullish flow, identify what rule produces that classification. When the value drives your entire thesis, compare it with an original record or another appropriate source where possible.
Try this on one hypothetical outlier: write its identity, time, units, coverage, adjustment basis, missing-data status, and definition. If any item remains unknown, say how that affects your confidence.
DAXA's research workflow is strongest when careful input checks sit alongside the interpretation. Explore the platform with a specific question and a data-quality checklist in hand.
Sources and further reading
Bring more context to your research
Explore DAXA’s research workspace and choose a membership that fits the information you use.