Design Notes
Why these charts
We kept it simple on purpose: show trends, compare pages fairly, and surface problems fast.
Questions we cared about
- Traffic When do we spike? (pageviews by day)
- Reliability Are errors trending up or down? (error rate)
- Perf vs Value Which slow pages are also popular?
- Reality check What happened recently? (activity table)
Chart choices
- Line β time trends (pageviews, error rate)
- Bar β category comparisons (per-page medians, top pages)
- Scatter β relationships (load vs pageviews)
- Grid β raw details for verification
What we measured
- Pageviews (14d) β time series β line
- Error rate (14d) β errors Γ· total β line (y starts at 0%)
- Median load by page β robust vs outliers β bar (Top 10)
- Load vs views β prioritize slow + popular β scatter
- Recent activity β quick sanity table (latest 100)
Skipped on purpose
No pie charts β theyβre hard to read precisely. We stuck to charts that decode quickly.
Deep-dives