Overview of a saved library
A chart on its own is decoration. Every chart here answers a question the user would otherwise have to ask, and you say the answer in words next to it.
Get the data
From the Bookmarks table via query_table. You need the whole table, not
a page — read total from the first call and page with offset until you
have it. Counting 100 of 4,000 saves and calling it a distribution is worse
than not charting at all.
Columns you'll use: Saved (ISO date), Topics (array), Site, Type,
Status.
The four that earn their place
- Saves over time — by week or month. Answers "am I still using this?" and shows bursts, which usually map to a project.
- Topics by volume — horizontal bars, biggest first. Answers "what am I actually collecting?" Cap at the top 12 and group the tail as "other"; a bar chart with 60 labels communicates nothing.
- Sources — top domains. Answers "where does my reading come from?" Often uncomfortable and therefore useful.
- Topic over time — the top few topics as lines. Answers "what am I moving toward, and what have I dropped?"
Skip any chart the data can't support. Fewer than ~20 saves, or fewer than three distinct topics, means the honest output is a sentence, not a chart.
Say what it means
Under each chart, one or two sentences of what it shows — the actual reading, not a caption. "Two thirds of your saves are from three domains" is worth writing. "This chart shows saves by domain" is not.
Close with what stands out across all four: a topic that stopped, a source that dominates, a burst that maps to something they were building.
Rules
- Charts follow the repo's dataviz conventions. Render as an HTML page in the user's stash so they can keep it.
- Label axes and units. Never a pie chart with more than five slices.
- Say the corpus size and date range you charted, so the reader knows what they're looking at.
