Merging several data sources into one chart hides a string of alignment decisions readers never see

What it is

Merging multiple data sources into one chart (internal tracking + third-party statistics + survey data) implicitly involves a series of alignment decisions: how the sources' definitions are unified, how time granularities align (daily-updated and monthly-updated data on one time axis), how conflicting values are resolved (two sources disagree on the same metric—which is shown), and how periods lacking a source are handled. Every one of these decisions shapes the final graphic's form, yet all are usually invisible—readers see a smooth merged curve with no knowledge of what processing happened at source transitions. Undeclared alignment makes a merged chart's credibility impossible to assess.

Why it happens

Alignment decisions shape conclusions because each is an implicit assumption. Definition unification: putting "internally counted active users" and "third-party SDK active devices" into one series implicitly claims the two are approximately equivalent—yet their difference may be systematic (devices outnumber users), and the slope change across the merged series' source segments gets misread as business trend. Time alignment: when daily and monthly data merge, how the monthly value renders during the first half of its month (carry-forward, gap, interpolation) determines the curve's shape at month boundaries. Conflict resolution: the disagreement between sources on the same metric may itself be valuable information (different statistical boundaries), and silently choosing one source discards that difference and may pick the more biased one. All these decisions can be made reasonably—the problem is that undeclared, readers cannot fold them into their reading of the graphic: every inflection point on the merged curve might be a data feature or the side effect of an alignment decision, indistinguishable on the chart. Declaring the alignment (a caption listing each source's coverage and processing rules) turns these side effects from invisible into assessable.

Where it stops holding

The declaration requirement scales with the depth of merging: light merging—one primary source with another as spot verification—needs only primary and verification sources labeled; full merging—splicing several sources along the time axis—needs each interval's source and processing rule stated individually. Alternatives to merging are sometimes better than merging: rendering multi-source data side by side (separate lines on one time axis, unspliced) preserves the differences at the cost of readers integrating them themselves—for professional readers this is the more honest option. Technical feasibility also has a boundary: sources with fundamentally incompatible definitions (different statistical objects) cannot be truly merged no matter how the alignment is documented, and the honest choice there is to abandon the merge rather than force alignment.

Applying it

• State in multi-source chart captions each source's coverage interval, definition, and the merge processing rule (carry-forward / interpolation / gap).
• Place a vertical marker at source-transition points on the time axis, reminding readers that continuity there is manufactured.
• When conflicting values coexist, prefer rendering them side by side (one line each) with the difference annotated; silent resolution only where the difference is negligible.
• Verification: ask a reader to point out where "source A's interval" ends and "source B's" begins on the merged curve; if they cannot and no marker exists, alignment disclosure is missing.

Related

• Same group: U8.07.1 Definition changes create spurious trends · U8.07.2 Data source and update time must accompany the chart
• Nearby: U8.07.1 Definition changes create spurious trends · U8.07.2 Data source and update time must accompany the chart
• Search terms: data integration · source alignment · multi-source chart