Time Series Trend Analysis Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: badcd81d-3b9d-4d60-9143-156df7b5302f

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Analyze the time series in {series_context} for genuine trend, distinguishing it from the patterns that fool the eye. (1) Plot and describe the raw series first — level, apparent direction, volatility, and any obvious breaks — before fitting anything, because the eye catches structure that automated fits miss and vice versa. (2) Decompose into trend, seasonality, and remainder: identify and remove seasonal cycles (whose period you justify from the domain, not from fishing), since unremoved seasonality masquerades as trend or masks it. (3) Test the trend against autocorrelation: adjacent points in time are correlated, so ordinary significance tests overstate confidence — a rising run can be pure random walk. State how you accounted for this. (4) Check for structural breaks and regime changes rather than forcing one trend across a series that has clearly shifted. (5) Distinguish the mean-reverting from the persistent: a series that wanders (unit root) has no stable trend to extrapolate, which changes forecasting entirely. (6) State the conclusion — trend present/absent, its rate with uncertainty — and explicitly bound how far forward it can be projected before the analysis stops supporting it.

Analyze the time series in {series_context} for genuine trend, distinguishing it from the patterns that fool the eye. (1) Plot and describe the raw series first — level, apparent direction, volatility, and any obvious breaks — before fitting anything, because the eye catches structure that automated fits miss and vice versa. (2) Decompose into trend, seasonality, and remainder: identify and remove seasonal cycles (whose period you justify from the domain, not from fishing), since unremoved seasonality masquerades as trend or masks it. (3) Test the trend against autocorrelation: adjacent points in time are correlated, so ordinary significance tests overstate confidence — a rising run can be pure random walk. State how you accounted for this. (4) Check for structural breaks and regime changes rather than forcing one trend across a series that has clearly shifted. (5) Distinguish the mean-reverting from the persistent: a series that wanders (unit root) has no stable trend to extrapolate, which changes forecasting entirely. (6) State the conclusion — trend present/absent, its rate with uncertainty — and explicitly bound how far forward it can be projected before the analysis stops supporting it.

Variables

Name	Type	Required	Trust level
series_context		yes

time-seriestrend-detectionseasonalitychain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18