An interactive tool for exploring any count time series before you model it. Upload a spreadsheet of periodic counts like sales, cases, incidents, traffic, defects, anything counted over time and read each series' long-run trend, seasonal rhythm, overdispersion, and how multiple series move against one another, across whatever time resolutions your file contains. Everything runs locally in your browser; no data leaves your machine.
Set an exact date range below, or drag across the lower strip. The shaded band marks the COVID onset (early 2020).
Average count by position in the year — the cycle, isolated from trend.
Histogram of counts, with the variance-to-mean ratio that decides Poisson vs negative-binomial modelling.
Every series at a glance — sort to see which are largest, most volatile, most zero-prone, or most strongly trending.