Obsidian Heatmap

Obsidian heatmaps

Two complementary tools cover almost every heatmap need in Obsidian:

1. Heatmap Calendar plugin (Richardsl) — GitHub-style year overview, one call from a dataviewjs block. Handles 365-day grid, month labels, weekday rows, today border, color gradients, click/hover. Use this first when the question is "did I do X on each day of the year".

2. Custom DataviewJS heatmaps — anything the plugin cannot draw: weeks × 7 days for the last N weeks, rows × cols matrix (e.g. metrics × weeks), rows × N days (projects × dates), custom score functions, multi-metric tooltips, streaks. The plugin draws a year; custom DOM draws whatever shape you need.

Pick the plugin first; drop to custom only when the year frame is wrong or you need a non-calendar shape.

Decision: which heatmap

If the user shows a GitHub contributions screenshot — that is the last 53 weeks scrolling shape, not a calendar year. The plugin always renders Jan 1 → Dec 31 of year. For a rolling-N-weeks contribution graph use the custom pattern.

Part A — Heatmap Calendar plugin (Richardsl)

Repo: Richardsl/heatmap-calendar-obsidian. Install via Community plugins in Obsidian. Exposes a single global function on plugin load.

API surface

renderHeatmapCalendar(container, calendarData)

container is the DOM element to render into — from a dataviewjs block use this.container. The function is on window, so just call renderHeatmapCalendar(...).

calendarData fields

Entry object

{
  date: "YYYY-MM-DD",   // required, must be ISO; daily-note filenames work directly
  intensity: 3,         // optional, falls back to defaultEntryIntensity
  content: "🏋️",       // optional, rendered inside the cell (emoji, text, dv.span(...))
  color: "green"        // optional, key into colors object; defaults to first key
}

Color palettes

Each palette is an array of HEX or rgba() strings — one per intensity level. More entries = finer gradient.

colors: {
  green: ["#c6e48b", "#7bc96f", "#49af5d", "#2e8840", "#196127"],   // GitHub default
  blue:  ["#8cb9ff", "#69a3ff", "#428bff", "#1872ff", "#0058e2"],
  red:   ["#ffb3b3", "#ff7373", "#ff3333", "#cc0000", "#800000"],
}

Two ways to pass:

• Inline — pass the object directly in calendarData.colors. Self-contained, lives in the note.
• Named — define the palette once in Settings → Heatmap Calendar, then pass colors: "myPalette" (string). Reusable across notes; survives moving the block.

How intensity → color level works: plugin scales each entry's intensity from [intensityScaleStart, intensityScaleEnd] to [1, palette.length], rounds, and picks the corresponding HEX. So a palette of 5 colors gives 5 visible levels regardless of raw values.

CSS hooks

The plugin renders a known DOM tree you can style via Settings → Appearance → CSS snippets:

Cell background-color is inlined — to override, use !important in your snippet.

Minimal example

const data = {
  year: 2026,
  entries: dv.pages('"Daily"')
    .where(p => p.exercise)
    .map(p => ({
      date: p.file.name,           // daily-note name = "YYYY-MM-DD"
      intensity: p.exercise,
      content: "🏋️",
    }))
    .array(),
};
renderHeatmapCalendar(this.container, data);

Multi-color overlay

Two metrics on the same year — entries with the same date can both render (last one wins per cell, so prefer disjoint days or accept overwrite):

const data = {
  year: 2026,
  colors: {
    green: ["#c6e48b", "#7bc96f", "#49af5d", "#2e8840", "#196127"],
    blue:  ["#8cb9ff", "#69a3ff", "#428bff", "#1872ff", "#0058e2"],
  },
  intensityScaleStart: 1,
  intensityScaleEnd: 60,
  entries: [],
};

for (const p of dv.pages('"Daily"').where(p => p.exercise)) {
  data.entries.push({ date: p.file.name, intensity: p.exercise, content: "🏋️", color: "green" });
}
for (const p of dv.pages('"Daily"').where(p => p.meditation)) {
  data.entries.push({ date: p.file.name, intensity: p.meditation, content: "🧘", color: "blue" });
}

renderHeatmapCalendar(this.container, data);

When two metrics matter equally per day, don't overlay — split into two heatmap blocks (one per metric) stacked vertically. Overlay is for highlighting outliers, not comparison.

Pick a data source first

The plugin doesn't care where data comes from — but the choice of source shapes everything else (tracking friction, edit ergonomics, what queries are cheap). Pick deliberately:

These are not exclusive — you can run two heatmap blocks side by side, one driven by frontmatter and another by tags. Don't pick more than one source for the same metric in the same heatmap; mixed sources produce inconsistent results when the same date is recorded twice.

Pattern — entries from frontmatter

Daily-note frontmatter is the cleanest source: one note per day, named YYYY-MM-DD.md, frontmatter has exercise: 3 etc.

renderHeatmapCalendar(this.container, {
  year: dv.date("today").year,
  entries: dv.pages('"Daily"')
    .where(p => p.workout != null)
    .map(p => ({ date: p.file.name, intensity: p.workout, content: "🏋️" }))
    .array(),
});

Pattern — entries from inline-fields in monthly logs

When you don't keep a daily note per day but log inside a monthly file using flat bullets — for example, a reading log:

- 2026-04-25 [pages:: 32] [minutes:: 45]
- 2026-04-26 [pages:: 18] [minutes:: 25]
- 2026-04-27 [pages:: 60] [minutes:: 90]

Dataview parses these into list items with date in text and inline-fields as keys. Score them and render:

// Pages is the goal; minutes adds a small bonus so a long careful read still registers
const score = e => (e.pages ?? 0) + (e.minutes ?? 0) * 0.3;

const entries = dv.pages('"Logs/Monthly"')
  .file.lists
  .where(l => /^\d{4}-\d{2}-\d{2}/.test(l.text))
  .map(l => {
    const date = l.text.match(/^\d{4}-\d{2}-\d{2}/)[0];
    return { date, intensity: score(l), content: "📖" };
  })
  .array();

renderHeatmapCalendar(this.container, { year: dv.date("today").year, entries });

Adapt the schema to whatever you track — running (km, minutes), writing (words, sessions), language study (new_words, minutes). The score function is where you decide what the cell intensity means: pure output, time-weighted, or composite.

Pattern — entries from task counts

Cell intensity = number of tasks completed on that day, from ✅ YYYY-MM-DD markers in any task:

const byDate = new Map();
for (const t of dv.pages().file.tasks.where(t => t.completed && t.completion)) {
  const d = t.completion.toString().slice(0, 10);
  byDate.set(d, (byDate.get(d) ?? 0) + 1);
}
const entries = Array.from(byDate, ([date, n]) => ({ date, intensity: n, content: "" }));
renderHeatmapCalendar(this.container, { year: dv.date("today").year, entries });

Pattern — entries from tag presence

Two flavors, depending on what "presence" means.

Binary habit on daily notes — daily note named YYYY-MM-DD.md has tag #workout → that day is "done":

const TAG = "#workout";
const entries = dv.pages('"Daily"')
  .where(p => p.file.tags.includes(TAG))
  .map(p => ({ date: p.file.name, intensity: 1, content: "🏋️" }))
  .array();
renderHeatmapCalendar(this.container, {
  year: dv.date("today").year,
  entries,
  intensityScaleStart: 0,
  intensityScaleEnd: 1,    // binary — every "done" day uses the top color
});

p.file.tags includes nested tags too (#workout/morning matches #workout). For exact-match only, compare manually: p.file.tags.some(t => t === TAG).

Count of tagged pages per day — intensity = number of pages with #read whose creation date falls on day Y. Works without per-day notes:

const TAG = "#read";
const byDate = new Map();
for (const p of dv.pages(TAG)) {
  const d = p.file.cday.toFormat("yyyy-MM-dd");
  byDate.set(d, (byDate.get(d) ?? 0) + 1);
}
const entries = Array.from(byDate, ([date, n]) => ({ date, intensity: n, content: "📖" }));
renderHeatmapCalendar(this.container, { year: dv.date("today").year, entries });

p.file.cday is creation day; p.file.mday is last-modified day. For "what did I read today" use cday; for "what did I touch today" use mday. If your note has an explicit date: frontmatter, prefer that — file timestamps reflect filesystem operations, not the event the note describes.

Multi-tag overlay — one heatmap, many tags, color per tag:

const TAGS = [
  { tag: "#workout",    color: "orange", emoji: "🏋️" },
  { tag: "#meditation", color: "blue",   emoji: "🧘" },
  { tag: "#read",       color: "green",  emoji: "📖" },
];
const entries = [];
for (const { tag, color, emoji } of TAGS) {
  for (const p of dv.pages('"Daily"').where(p => p.file.tags.includes(tag))) {
    entries.push({ date: p.file.name, intensity: 1, content: emoji, color });
  }
}
renderHeatmapCalendar(this.container, {
  year: dv.date("today").year,
  colors: {
    orange: ["#ffa244", "#fd7f00"],
    blue:   ["#8cb9ff", "#428bff"],
    green:  ["#7bc96f", "#196127"],
  },
  entries,
});

Same caveat as multi-color overlay: when two tags hit the same day, the last entry in the array wins the cell. Order TAGS so the more important tag goes last, or split into separate heatmaps.

Caveats with the plugin

• One year per call. year filters entries; cross-year data renders as empty in the wrong year. Stack two blocks for two years.
• renderHeatmapCalendar may be undefined on first paint if the plugin loads after the dataview block. Guard with if (typeof renderHeatmapCalendar === 'function') {...} and reload the note. Persistent issue → set "Show ribbon icon" off and on, or restart Obsidian.
• Daily-note name must be YYYY-MM-DD. If your daily notes use a different format, derive the date string explicitly (p.date from frontmatter, or dv.func.dateformat(p.file.day, "yyyy-MM-dd")).
• Entries with the same date overwrite. Last entry in array wins. If you build entries in a loop, latest insert is what renders.
• content is a string, not HTML. For interactive content (links to the daily note), pass content: await dv.span(\[[${p.file.name}]]`)` and the dataview span renders inline.
• Hover preview needs Page Preview enabled. Core plugin → Page Preview → toggle "Source mode" / "Reading view" depending on where the heatmap renders.
• Color array length determines resolution. 3 colors → 3 visible levels; 5 → 5; 10 → smooth gradient. Don't pass 100 — eyes can't tell.
• intensityScaleStart === intensityScaleEnd (or all entries have the same intensity) makes every cell the top color. Set explicit start/end if your data is bimodal.
• separateMonths: true breaks the visual rhythm of the GitHub layout — only enable if you want per-month dividers for styling.

When to override colors via CSS instead of colors field

Use colors field when each block has its own palette. Use a CSS snippet when:

• You want the same palette across the entire vault and tire of repeating it.
• You want CSS variables driven by the active theme (light/dark).
• You want to tweak .isEmpty background (the inline background-color on .hasData overrides palette, but .isEmpty has only the CSS class, easily themable).

/* CSS snippet — empty cells barely visible in dark theme */
.theme-dark .heatmap-calendar-boxes li.isEmpty {
  background-color: rgba(148, 163, 184, 0.06);
}
.heatmap-calendar-boxes li.today {
  outline: 1.5px solid var(--interactive-accent) !important;
}

Part B — Custom DataviewJS heatmaps

When the year-frame doesn't fit. Three reusable shapes cover most needs.

Pattern B1 — weeks × 7 days, last N weeks (GitHub-contribution shape)

This is what most "GitHub-style" requests actually want: rolling N weeks, scroll cuts off naturally, today on the right.

const WEEKS = 14;
const MAX = 60;                          // 60 pages saturates the color
const score = e => e.pages ?? 0;

// Source: flat bullets `- YYYY-MM-DD [pages:: N] [minutes:: M]`
const entries = [];
for (const l of dv.pages('"Logs/Monthly"').file.lists) {
  const m = l.text.match(/^(\d{4}-\d{2}-\d{2})/);
  if (!m) continue;
  entries.push({ date: m[1], pages: l.pages ?? 0, minutes: l.minutes ?? 0 });
}
const byDate = new Map(entries.map(e => [e.date, e]));

const todayDt = dv.date("today");
const start = todayDt.startOf("week").minus({ weeks: WEEKS - 1 });
const wrap = this.container.createEl("div", { attr: { style: "overflow-x: auto; padding: 8px 0;" } });
const flex = wrap.createEl("div", { attr: { style: "display: flex; gap: 3px; align-items: flex-start;" } });

// Day-of-week label column (Mon, Wed, Fri)
const dayCol = flex.createEl("div", { attr: { style: "display: flex; flex-direction: column; gap: 3px; padding-top: 16px;" } });
for (const dl of ["Mon", "", "Wed", "", "Fri", "", "Sun"]) {
  dayCol.createEl("div", { text: dl, attr: { style: "height: 16px; font-size: 9px; color: #64748b; line-height: 16px;" } });
}

// Week columns
for (let w = 0; w < WEEKS; w++) {
  const col = flex.createEl("div", { attr: { style: "display: flex; flex-direction: column; gap: 3px;" } });
  const weekStart = start.plus({ weeks: w });
  // Show month label on the column when first day of month falls in this week
  const monLabel = weekStart.day <= 7 ? weekStart.toFormat("LLL") : "";
  col.createEl("div", { text: monLabel, attr: { style: "height: 13px; font-size: 9px; color: #94a3b8; text-align: center;" } });

  for (let d = 0; d < 7; d++) {
    const cellDate = weekStart.plus({ days: d });
    const ds = cellDate.toFormat("yyyy-MM-dd");
    const data = byDate.get(ds);
    const isFuture = cellDate > todayDt;
    const isToday = ds === todayDt.toFormat("yyyy-MM-dd");

    let bg, title;
    if (isFuture) {
      bg = "rgba(148,163,184,0.04)";
      title = `${ds}: future`;
    } else if (data) {
      const sc = score(data);
      const intensity = Math.min(1, sc / MAX);
      const light = 25 + Math.round(intensity * 30);
      bg = `hsl(120, 65%, ${light}%)`;
      title = `${ds}: score ${sc}`;
    } else {
      bg = "rgba(239,68,68,0.10)";
      title = `${ds}: missed`;
    }
    const border = isToday ? "1.5px solid #3b82f6" : "1px solid transparent";
    col.createEl("div", { attr: { style: `width: 16px; height: 16px; background: ${bg}; border: ${border}; border-radius: 3px;`, title } });
  }
}

Knobs to tune: WEEKS, MAX, the score function, the empty/missed colors. Future cells stay barely visible; today gets a blue outline; missed cells use a soft red so gaps are visible at a glance.

Pattern B2 — rows × cols matrix (metrics × weeks)

For "how does each component look across the last 8 weeks". Cell text shows the value; color encodes ratio to row-specific max.

// data shape:
//   rows = [{ key, label, max }]
//   cols = [{ key, label }]
//   value(rowKey, colKey) → number | null

// Habit tracker: rows = habits (each with its own weekly target), cols = recent weeks.
// Cell value = number of times that habit happened in that week.
const rows = [
  { key: "exercise",   label: "Exercise",   max: 5 },   // up to 5 sessions/wk (cap to avoid overtraining)
  { key: "reading",    label: "Reading",    max: 7 },   // daily target
  { key: "journaling", label: "Journaling", max: 7 },   // daily target
  { key: "language",   label: "Language",   max: 7 },   // daily target
];
const cols = [
  { key: "2026-W18", label: "W18" },
  { key: "2026-W17", label: "W17" },
  { key: "2026-W16", label: "W16" },
];
const data = {
  "2026-W18": { exercise: 4, reading: 6, journaling: 5, language: 7 },
  "2026-W17": { exercise: 5, reading: 7, journaling: 4, language: 6 },
  "2026-W16": { exercise: 3, reading: 5, journaling: 6, language: 4 },
};
const value = (rk, ck) => data[ck]?.[rk] ?? null;

const wrap = this.container.createEl("div", { attr: { style: "overflow-x: auto; padding: 8px 0;" } });
const grid = wrap.createEl("div", {
  attr: { style: `display: grid; grid-template-columns: 90px repeat(${cols.length}, minmax(56px, 1fr)); gap: 3px; font-size: 12px;` }
});

// header
grid.createEl("div", { text: "" });
for (const c of cols) {
  grid.createEl("div", { text: c.label, attr: { style: "color: #94a3b8; text-align: center; font-size: 10px; align-self: end;" } });
}

// rows
for (const r of rows) {
  grid.createEl("div", { text: r.label, attr: { style: "color: #cbd5e1; font-weight: 500; padding: 6px 4px; align-self: center;" } });
  for (const c of cols) {
    const v = value(r.key, c.key);
    const ratio = v == null ? null : Math.max(0, Math.min(1, v / r.max));
    let bg = "rgba(148,163,184,0.1)";
    let color = "#64748b";
    if (ratio != null) {
      const hue = Math.round(ratio * 120);          // 0 = red, 120 = green
      const light = 28 + Math.round(ratio * 22);    // 28% → 50%
      bg = `hsl(${hue}, 65%, ${light}%)`;
      color = ratio > 0.4 ? "white" : "#e2e8f0";
    }
    const cell = grid.createEl("div", {
      attr: {
        style: `background: ${bg}; border-radius: 4px; padding: 10px 4px; text-align: center; color: ${color}; font-weight: 600;`,
        title: `${r.label} ${c.label}: ${v ?? "—"}/${r.max}`,
      },
    });
    cell.textContent = v == null ? "—" : (Number.isInteger(v) ? v : v.toFixed(1));
  }
}

Use this when each row has its own max — habits with different weekly ceilings (exercise capped at 5/wk, reading at 7/wk), or score components that each cap at a different number. A single global max forces low-cap rows to always look saturated and high-cap rows to always look pale, hiding the signal.

Pattern B3 — rows × N days (project activity, sticky labels)

Many rows (10+ projects), many columns (60 days), labels must stay visible while scrolling sideways.

const days = 60;
const cellSize = 11;
const today = dv.date("today").startOf("day");

// Source: rows = [{ name, files: [{ mtime }] }] — files anywhere with luxon mtime
const rows = [
  { name: "Project A", files: dv.pages('"Project A"').array().map(p => ({ mtime: p.file.mtime })) },
  { name: "Project B", files: dv.pages('"Project B"').array().map(p => ({ mtime: p.file.mtime })) },
].filter(r => r.files.length);

const stickyBg = "var(--background-primary)";
const labelStyle = `font-weight: 600; color: var(--text-normal); font-size: 11px; white-space: nowrap; position: sticky; left: 0; background: ${stickyBg}; z-index: 1; padding: 2px 6px 2px 0;`;
const cornerStyle = `position: sticky; left: 0; background: ${stickyBg}; z-index: 1;`;

const scroller = this.container.createEl("div", { attr: { style: "overflow-x: auto; padding-bottom: 4px;" } });
const grid = scroller.createEl("div", {
  attr: { style: `display: grid; grid-template-columns: 130px repeat(${days}, ${cellSize}px); gap: 2px; align-items: center;` }
});

// Header — month labels on the 1st of each month
grid.createEl("div", { attr: { style: cornerStyle } });
for (let d = 0; d < days; d++) {
  const date = today.minus({ days: days - 1 - d });
  const showLabel = date.day === 1 || d === 0;
  grid.createEl("div", { text: showLabel ? date.toFormat("LLL") : "", attr: { style: "font-size: 9px; color: var(--text-muted);" } });
}

// Data rows — bucket files into days, intensity = count
for (const r of rows) {
  const buckets = new Array(days).fill(0);
  for (const f of r.files) {
    const fday = dv.date(f.mtime).startOf("day");
    const diff = Math.floor(today.diff(fday, "days").days);
    if (diff >= 0 && diff < days) buckets[days - 1 - diff] += 1;
  }
  const max = Math.max(...buckets, 1);

  grid.createEl("div", { text: r.name, attr: { style: labelStyle } });
  for (let d = 0; d < days; d++) {
    const v = buckets[d];
    const intensity = v === 0 ? 0 : 0.25 + (v / max) * 0.75;
    const bg = v === 0 ? "rgba(148,163,184,0.07)" : `rgba(34,197,94,${intensity})`;
    grid.createEl("div", {
      attr: { style: `width: ${cellSize}px; height: ${cellSize}px; background: ${bg}; border-radius: 2px;`, title: v ? `${v} files` : "" }
    });
  }
}

// Auto-scroll to today on the right edge
requestAnimationFrame(() => { scroller.scrollLeft = scroller.scrollWidth; });

The position: sticky on row labels keeps them visible while the cells scroll horizontally — without it, labels disappear and you can't tell which row is which past day 30.

Aux helpers

Current streak

Days-in-a-row done, counted from today (or yesterday if today not yet recorded) backwards.

const currentStreak = (entries) => {
  const todayDt = dv.date("today");
  const todayDs = todayDt.toFormat("yyyy-MM-dd");
  const lookup = new Map(entries.map(e => [e.date, e]));
  let cursor = lookup.has(todayDs) ? todayDt : todayDt.minus({ days: 1 });
  let streak = 0;
  while (true) {
    const ds = cursor.toFormat("yyyy-MM-dd");
    const data = lookup.get(ds);
    if (data && data.done) { streak++; cursor = cursor.minus({ days: 1 }); }
    else break;
  }
  return streak;
};

done can be derived: data.done = (data.score > 0) if the user doesn't track an explicit done flag.

Missed days

Days with no entry between first recorded date and yesterday.

const missedDays = (entries) => {
  if (!entries.length) return 0;
  const todayDs = dv.date("today").toFormat("yyyy-MM-dd");
  const lookup = new Set(entries.map(e => e.date));
  let missed = 0;
  let cursor = dv.date(entries[0].date);
  while (cursor.toFormat("yyyy-MM-dd") < todayDs) {
    if (!lookup.has(cursor.toFormat("yyyy-MM-dd"))) missed++;
    cursor = cursor.plus({ days: 1 });
  }
  return missed;
};

Best practices

1. Default to the plugin. Year overview, the most common request, is one block of code with the plugin and 60 lines without. Drop to custom only when the shape genuinely doesn't fit.

2. One metric per block. Tempting to overlay everything; in practice readers can't decode multi-color cells. Stack two heatmaps over each other instead.

3. Show empty days deliberately. GitHub uses a barely-visible gray for empty-but-tracked days. Use a softer gray for "before tracking started" and a soft red for "missed" — the contrast between empty and missed is what makes streaks legible.

4. Today border, not today fill. If you fill today with the strongest color, the user can't read its intensity. Outline or border instead.

5. Cell size matters at scale. 11px cells give 53 weeks × 7 days a width of ~650px. 16px gives a friendlier shape but only ~30 weeks fit before scroll. Match cell size to viewport and how often the heatmap is reviewed.

6. title attribute is your tooltip. No need for a tooltip library — title="..." on each cell gives free hover text in any browser. Keep titles compact: 2026-04-25: 32 pages, 45 min not multi-line.

7. Color hue from data, lightness from intensity. hsl(${hue}, 65%, ${light}%) lets you compare cells across rows with different scales. Pure red→green is intuitive for "bad → good"; if direction is ambiguous, stick to a single hue and vary lightness only.

8. Future cells barely visible. Don't render them as "missed" — they haven't happened yet. Use rgba(148,163,184,0.04) (almost transparent) so the grid shape stays but the user's eye doesn't read failure.

9. Auto-scroll right on rolling-window heatmaps so today is visible on first render. requestAnimationFrame(() => { scroller.scrollLeft = scroller.scrollWidth; }) is the one-line fix.

10. Sticky labels for rows × N days. Without position: sticky the row name disappears as soon as you scroll horizontally. Always pair sticky-left labels with a sticky-left "corner" cell on the header row.

When NOT to use a heatmap

• You have ≤ 7 data points. A bar chart or sparkline shows them better.
• You care about absolute values, not patterns. A table with sortable columns is more honest.
• The metric is monotonically increasing. Cumulative-progress charts (line, area) communicate growth; heatmap won't.
• Most days are zeros. Heatmaps are pattern-detectors; if 95% of cells are empty, the eye sees noise.
• You want to compare two people / two metrics on equal footing. Side-by-side bars or a scatter plot answers "are they correlated" better than two stacked heatmaps.

Common gotchas

• Plugin shows nothing — likely the daily-note name isn't YYYY-MM-DD, or entries array is empty after filtering by year, or renderHeatmapCalendar isn't loaded yet (reload the note).
• All cells same color — intensity is the same value for every entry, or all entries fall outside [intensityScaleStart, intensityScaleEnd] and clamp to one bound.
• Today's cell on the wrong day — Obsidian's dv.date("today") uses local time; the plugin uses UTC for cross-year filter and local for the today-border. In timezones far from UTC the off-by-one shows around midnight; rarely matters in practice.
• CSS snippet doesn't apply — plugin sets style="background-color: ..." inline on .hasData cells; use !important to override, or target .isEmpty only (no inline style there).
• Custom heatmap renders empty — dv.pages(...) returned a Proxy, not an array. Call .array() before .map() on Dataview Proxies if you intend to use plain array methods.
• Block flickers on scroll — if your DataviewJS heatmap is far down a long note, Obsidian re-renders it lazily on scroll. Keep heavy work outside the render loop, or move large logic into a plugin / shared script and call a thin wrapper from the block.