# long-term Habit Tracker

**URL:** <https://forum.obsidian.md/t/long-term-habit-tracker/118614>\
**Category:** Help\
**Created:** [September 22, 2026, 10:53am UTC](https://forum.obsidian.md/t/long-term-habit-tracker/118614 "2026-09-22T10:53:55Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![davidlearneveryday](https://forum.obsidian.md/letter_avatar_proxy/v4/letter/d/7c8e57/32.png) [@davidlearneveryday](https://forum.obsidian.md/u/davidlearneveryday)\
**Post date:** [September 22, 2026, 10:53am UTC](https://forum.obsidian.md/t/long-term-habit-tracker/118614/1 "2026-09-22T10:53:55Z")

</div>

> First, search the [help docs](https://help.obsidian.md) and this forum. Maybe your question has been answered! The [debugging steps](https://forum.obsidian.md/t/about-the-bug-reports-category/24/11) can help, too. Still stuck? Delete this line and proceed.

### Hi everyone,

I’m currently evaluating different architectures to set up a habit tracker in my vault. My primary goals are long-term maintainability, performance, and keeping my data as native to Markdown as possible.

The setup I’m looking into consists of:

- **Data layer:** Storing habit status using boolean/checkbox properties in the YAML frontmatter of my daily notes (`habit_name: true/false`).

- **Dashboard:** A single Canvas file containing individual cards per habit.

- **Visualization / Logic:** Using `dataviewjs` code blocks inside each card to read the daily notes, calculate streaks/completion rates, and render visual monthly calendars.

On paper, this seems like a solid approach since it avoids single-purpose community plugins and keeps everything contextualized within daily notes. However, before committing to it, I’d love to get feedback from anyone running this or similar setups:

1. **Performance & Scalability:**  
As the vault grows over time (hundreds of daily notes), does executing multiple `dataviewjs` queries across Canvas cards lead to noticeable rendering lag or high memory usage, particularly on mobile?

2. **Maintenance vs. Dedicated Plugins:**  
Compared to dedicated plugins (like Habit Tracker 21, Heatmap Tracker, etc.) or native Obsidian database views, is maintaining custom `dataviewjs` blocks worth the extra flexibility in your experience?

3. **Bottlenecks:**  
Are there any unforeseen drawbacks or edge cases with this architecture that I should be aware of for long-term vault longevity?

I’d really appreciate any insights, alternative recommendations, or lessons learned from your own workflows!

### Things I have tried

---

<div class="post-metadata">

**Author:** ![uka.dev](https://forum.obsidian.md/user_avatar/forum.obsidian.md/uka.dev/32/96364_2.png) [@uka.dev](https://forum.obsidian.md/u/uka.dev)\
**Post date:** [September 28, 2026, 10:49am UTC](https://forum.obsidian.md/t/long-term-habit-tracker/118614/2 "2026-09-28T10:49:13Z")

</div>

The daily-note frontmatter part sounds like a durable source of truth. I would be more cautious about making the Canvas carry many separate `dataviewjs` cards as the vault grows.

Keep the daily notes and their boolean fields as the data layer, but start with one dashboard note and one or two queries that answer the questions you actually check. If the Canvas stays useful, let it be a view over those queries rather than the place where logic or history lives.

```dataview
TABLE habit_reading, habit_exercise
FROM "Daily"
WHERE file.day >= date(today) - dur(30 days)
SORT file.day DESC

```

I would also put the habit fields in a daily-note template so the schema stays consistent. The main trade-off is not Markdown versus plugins, it is how much custom query logic you want to maintain when a simple table no longer answers the question.

---

<div class="post-metadata">

**Author:** ![dimagious](https://forum.obsidian.md/user_avatar/forum.obsidian.md/dimagious/32/96654_2.png) [@dimagious](https://forum.obsidian.md/u/dimagious)\
**Post date:** [September 29, 2026, 4:02pm UTC](https://forum.obsidian.md/t/long-term-habit-tracker/118614/3 "2026-09-29T16:02:45Z")

</div>

Your data layer will age well: checkbox properties in daily notes are plain YAML, any tool can read them, and you can switch dashboards later without touching the data.

On your three questions:

1. Performance. Each dataviewjs block runs its own query over the vault and re-renders on its own, so the cost grows with the number of cards times the number of daily notes. Before committing, I’d build the Canvas with all the cards you plan and open it on your phone with a year of test notes.

2. Maintenance. Most of the time goes into edge cases in streak and calendar code: a missing note vs a note with `false`, days that shift across midnight or time zones, weeks starting on Sunday vs Monday, a habit you skip on weekends. You own each of those and have to test it again after Dataview or Obsidian updates.

3. Canvas. I’d keep one dashboard note and, if you still want the Canvas, embed that note as a card. The logic then lives in one place.

Disclosure: I wrote a plugin for the display layer, Dashy (in the community catalog). It reads the same checkbox properties from Obsidian’s metadata cache, so your data stays as you designed it, and the blocks are YAML with no JavaScript to maintain. One habit, with a folder called `Daily`:

```heatmap
source: Daily
field: habit_reading
range: month
title: Reading, this month

```

```stats
columns: 2
items:
  - { label: Longest reading streak, source: Daily, field: habit_reading, agg: streak, unit: " days" }
  - { label: Reading days this month, source: Daily, field: habit_reading, agg: sum, period: month }

```

```progress
items:
  - { label: Reading this month, source: Daily, field: habit_reading, agg: sum, period: month, goal: 30 }

```

Unticked days stay empty on the calendar, and the date comes from the note name (`2026-09-28.md`), so you do no time zone math. One limit: the streak card shows the longest run on record, and Dashy has no current-streak number yet. If today’s streak is the number you care about most, keep your own code for that card.

I haven’t benchmarked Dashy on a vault with thousands of notes, so the same advice applies: try it on your phone with your own data before you commit.
