The Slack Dashboard Bible

What is the Slack dashboard?

In February 2016 I set up a personal Slack workspace called "Mike's Links." Not for a team. Not for work. Just channels, each named after a topic, where I could drop things I wanted to remember. A recipe. An article about AI. A good tweet. A job I wanted to keep an eye on. The workspace became a personal media filing cabinet — 459 channels, one person, a decade of attention.

In April 2026 I ran a Python script across the full dataset for the first time. The script pulled every message from every channel via the Slack API, both for a 90-day live snapshot and year-by-year going back to 2016. The output was a dashboard and a series of visualisations that turned a decade of saved links into something readable.

The dashboard lives at mikelitman.me/slack-dashboard. It shows the 90-day live picture of the workspace — what I'm paying attention to right now — plus an all-time section showing how interest in each of the top 50 channels has moved across every year from 2016 to 2026.

The dashboard refreshes automatically every Sunday at 10pm via a cron job that re-runs the Python script, commits the updated data to the portfolio repo, and pushes to Netlify. The all-time phases data is updated at the same time.

All content: decks, posts, dashboard

The Slack dashboard project has generated more content than anything else in the portfolio. Here is everything, in order:

How it was built

The entire stack is open and repeatable. No proprietary tools, no external services beyond the Slack API itself.

The script: slack-full-analysis.py, a Python file that lives at ~/slack-full-analysis.py. It does two things in sequence. First, it runs a 90-day analysis: pulling every message from every channel active in the last 90 days, counting message volumes, extracting URLs, parsing timestamps, computing domain frequency and day-of-week distribution. Second (Step 5, added April 2026), it runs a year-by-year phase analysis: for each of the top 50 channels, it queries the Slack API once per year from 2016 to the current year, using Unix timestamp windows to isolate each calendar year. 550 individual API queries in total.

The dashboard: A single HTML file at deploy/slack-dashboard.html. It reads /data/slack-analysis-data.json (the 90-day data, baked in at build time) and /data/slack-phases-data.json (the all-time phases, fetched client-side). No framework, no build step — vanilla JS that renders the charts and tables from the raw JSON. The whole thing is one file.

The automation: A crontab job fires every Sunday at 10pm: it runs the Python script, then uses a git diff --cached --quiet || git commit pattern to commit only if the data has actually changed, then pushes to GitHub. GitHub Actions triggers a Netlify build. The full refresh cycle takes under 10 minutes.

The decks: Both the /slack deck and the /saves deck are pure HTML using Alpine.js for slide transitions. No PowerPoint. No export step. The bar charts are inline CSS width percentages derived directly from the data.

The blog posts: Written in the same HTML template as all portfolio blog posts, pulling from the same pages.css stylesheet. All data cited is drawn directly from slack-analysis-data.json and slack-phases-data.json — nothing estimated or invented.

Why it was built

The honest answer is curiosity. I had a decade of saves and no idea what the shape of it looked like. I knew I had a lot of channels. I didn't know how many were dormant, which ones had dominated each year, or what the aggregate arc looked like when you put 10 years of data in a single view.

The more interesting answer is that it was built to prove a thesis: that behaviour is more honest than opinion. I write about culture and media for a living. I can tell you what I think I care about. The data tells a different, more honest story. The gap between those two things is the interesting part.

There was also a practical motivation. The workspace had become invisible — I was adding to it every day without ever looking at the shape of what I was building. Running the analysis made the invisible visible. It turned a passive habit into an active portrait.

Finally: it's a portfolio piece. The dashboard demonstrates data pipeline thinking, API integration, front-end data visualisation, and the ability to build something that generates genuine insight from raw data. It shows both technical and analytical capability in a single artefact. That matters for the kind of roles I'm targeting.

The key data

Top 10 channels by all-time saves (2016–2026):

Top 10 channels by 90-day activity (Jan–Apr 2026):

A decade of findings

Running the year-by-year analysis across 50 channels from 2016 to 2026 surfaced a set of findings that are hard to arrive at any other way. These are not interpretations — they are visible in the raw numbers.

2022 was a dead year. Almost every channel simultaneously hit its lowest point. #ai: 7 saves (down from 41). #jobs: 23 (down from 204). #strategy: 14. #food: 5. There is no single explanation. The data shows a year of disengagement across the board — a year between identities.

The AI inflection is a visible cliff. #ai-artificial-intelligence went from 7 saves in 2022 to 581 in 2023 — an 83x jump triggered by the launch of ChatGPT in December 2022. No other channel shows anything close to this rate of change. 2024 was even higher at 896. The cliff is not a metaphor. It is literally visible in the bar chart.

#cultural-interface came from nowhere and became #1. Created in 2023, immediately the most active channel in the workspace, peaking at 1,253 saves in 2025 alone — more than any other channel has ever had in a single year. It didn't exist before 2023. It now defines the workspace. The thing that dominates your attention today may not have existed three years ago.

#jobs never sleeps. Active every single year from 2016 to 2026. Ten out of ten. The 2024 figure of 435 saves is the highest single-year figure of any channel in the entire workspace — higher even than #cultural-interface in its peak year, if you adjust for when it started. #jobs is the most honest channel in the workspace. It never lies about where my head is.

#strategy peaked in 2020 and never came back. 172 saves in 2020, 134 in 2021, then 14 in 2022. It has not recovered. This is the visible signature of moving from "strategist who studies the industry" to "founder who ships things." The decline is not noise. It is a clean break.

#food was a lockdown artefact. 17 saves in 2018, 68 in 2019, then 199 in 2020 during the pandemic, back to 69 in 2021, collapse to 5 in 2022. One external event, one year, one spike. A natural experiment in forced behaviour change.

#tools peaked in 2018 and has been declining for eight years. 82 saves at peak, around 24 a year now. This is the transition from being a user of other people's tools to being a maker of your own. The interest in collecting tools evaporated at exactly the point building started.

#events never broke. Active every single year from 2016 to 2026. From 2 saves in the first year to a peak of 68 in 2023. Through every career phase — agency, Web3, Contagious, founder — the interest in events persisted. It is the one unbroken thread.

#portfolio-page: nine years of the same question. Active since 2017, peaking at 92 saves in 2021, never reaching zero. The question of how to present yourself professionally is permanent. It just changes shape depending on what you're trying to present.

Every 2026 channel is a live product. #withmoshi-aivoice, #buggysmart, #claude-code, #pattern-media, #firstorder-aivoice, #thequeueindex. Zero new "interest area" channels created in 2026. The workspace's mental model has shifted entirely from curation to production.

#ff (fantasy football): #3 all-time. 1,951 total saves, ahead of #jobs (1,273) and #strategy (571). The 2024–25 Premier League season, when Nottingham Forest mounted their closest-ever challenge for a Champions League place, drove 712 saves in 2024 and 562 in 2025 — 1,274 saves across a single football season. This is the most honest number in the dataset. Nobody building a personal knowledge infrastructure wants to admit that football beats strategy. But the data doesn't lie.

#substack died and came back. 78 saves in 2021, collapsed to 7 in 2022, recovered to 57 in 2023, then hit a new all-time high of 107 in 2025. Some channels that go quiet are not dead. They are waiting for the world to catch up with the interest.

#wearables: appeared from nowhere in 2025. Zero saves from 2016 to 2024. Then 175 saves in 2025, making it the 16th highest channel by volume that year. A completely new interest area with no prior signal. The workspace does not predict — it records.

Wednesday is the peak saving day. Of the 2,456 timestamps analysed, Wednesday has the highest message volume at 516. Not Monday (a common assumption for a "productive" day) and not the weekend (another assumption for someone reading widely). Wednesday. 22:00 is the peak hour. The workspace is a late-night, mid-week habit.

51% of all links come from Twitter/X and LinkedIn. Twitter/X: 631 links (34% of total). LinkedIn: 320 links (17%). These are the platforms I openly criticise. They are also where more than half of every link worth saving originates. Behaviour and opinion are not the same thing.

The Spotify Wrapped comparison

Spotify Wrapped is the obvious reference point. Every December, Spotify releases a personalised annual review — your top artists, top songs, listening minutes, genre breakdown — presented as a shareable, beautifully designed experience. It became a cultural moment because it turned private behaviour into a public identity signal. People share their Wrapped not just because it's interesting but because it says something about who they are.

The Slack dashboard is doing something structurally similar but in a fundamentally different domain. Spotify Wrapped tells you about your entertainment consumption. The Slack dashboard tells you about your intellectual consumption — what you chose to pay attention to, not just what the algorithm served you.

There are four meaningful differences between this and Spotify Wrapped:

1. The data is self-generated, not platform-generated. Spotify collects your listening passively. Every save in this workspace is an active choice — a deliberate act of curation. The signal-to-noise ratio is higher because every data point required intention.

2. The time horizon is longer. Spotify Wrapped is annual. This dataset spans ten years. The value is not just "what were you into this year" but "how did you change" — which is a harder and more interesting question.

3. The categorisation is self-defined. Spotify's genres are assigned algorithmically. The channels here are named and organised by the person doing the saving. The taxonomy reflects the person's own mental model, which is itself data.

4. It is not shareable in the same way. Spotify Wrapped works as a social product because music taste is universally relatable. The Slack dashboard is more personal and more niche — the channels mean something to me but would mean nothing without context to most people. The blog posts and decks exist partly to translate that context for an outside audience.

The Wrapped comparison also points at the largest gap: presentation. Spotify Wrapped is mobile-first, animated, and designed for 10-second consumption. The Slack dashboard is desktop-first, data-dense, and designed for extended reading. The two decks (/slack and /saves) are the closest equivalent to the "wrapped" format — a curated highlight reel rather than the full data picture.

The market landscape

The Slack dashboard exists at the intersection of three adjacent markets: personal knowledge management (PKM), quantified self / personal analytics, and data storytelling tools. Here is a lay of the land.

Personal Knowledge Management (PKM)

PKM is the broad category of tools for capturing, organising, and retrieving information. The dominant players are Notion, Obsidian, Roam Research, Logseq, and Bear. All of these are primarily writing and note-taking tools with tagging and linking as their organising principle. None of them are built around URL curation specifically, and none offer any form of analytics or pattern analysis over the saved material.

Pocket (acquired by Mozilla, now being wound down), Instapaper, and Readwise are the closest to what this workspace does — they are URL-first save tools. Readwise in particular has moved towards structured knowledge extraction (Readwise Reader), but it focuses on reading and annotation, not on aggregated behavioural patterns across a catalogue of saves.

Quantified Self / Personal Analytics

This is a smaller, more niche market. Tools like Exist.io aggregate data from multiple apps (sleep, exercise, mood, productivity) and surface correlations. Gyroflow, Daylio, and similar journalling apps add reflection layers. RescueTime and Toggl track time and attention at the app level. None of these tools focus on intellectual consumption — what you were interested in, not how long you spent at a desk.

Data Storytelling

Spotify Wrapped is the cultural touchstone. Apple Music Replay is its lower-effort equivalent. Last.fm has done music listening statistics for 20+ years. Goodreads has reading statistics but they are basic and non-visual. GitHub does a year-in-code contribution graph. None of these operate in the "intellectual consumption" space beyond their specific domain.

Here is how the main players compare on the dimensions that matter for this project:

The gap is clear: no tool combines URL-level save data, self-defined categories, long-term behavioural analytics, and narrative output. The Slack dashboard is occupying a genuinely empty space — not because it was designed to, but because Slack was the tool that happened to capture all of this data by accident.

The vision

The immediate vision is exactly what exists: a living dashboard that turns the workspace into a readable portrait. Updated weekly. Growing in sophistication as more insight layers are added.

The larger vision is the underlying idea: that what you save is who you are, and that this data — when properly surfaced — is one of the most honest and revealing portraits of a person's intellectual life that exists. The workspace is not a productivity tool or a note-taking system. It is a decade-long diary written in links.

The vision for this as a product category — rather than just a personal project — is "Spotify Wrapped for your intellectual life." A tool that takes the links you save, the channels you organise them into, the time and frequency patterns of your saving behaviour, and turns them into a rich annual or ongoing portrait of who you're becoming. Not what you consumed passively (Netflix, Spotify, TikTok), but what you chose to pay attention to.

That product does not exist. The closest thing is this dashboard. The difference between the personal project and the product is multi-user support, a proper data ingestion layer (not everyone uses Slack), and a presentation layer built for shareability.

What could be added

Monetisation paths

The current project is not a commercial product — it is a personal tool and a portfolio piece. But if it were to become one, here are the credible paths:

1. The personal analytics SaaS. A tool that connects to users' Slack workspaces (or browser bookmarks, or Readwise accounts) and generates the same kind of dashboard automatically. Monthly subscription, £5–15/month. The target audience is people who already use PKM tools and think analytically about their learning. Small market, but high intent. The core challenge is data ingestion: not everyone uses Slack as a personal scrapbook, so multi-source support would be required from the start.

2. The annual Wrapped product. A paid report that generates a beautifully designed annual review of your intellectual life. One-time payment per year, £10–30. Lower friction than a subscription, higher shareability. Works as a marketing funnel for the ongoing SaaS. Spotify Wrapped is free because it drives Spotify subscriptions — this version would be the paid product itself.

3. The organisational intelligence product. The same analysis applied to team Slack workspaces rather than personal ones. What is the team actually paying attention to? How do stated priorities compare with actual channel behaviour? This is a genuine business analytics product and the market is significantly larger. It connects to existing "workplace analytics" tools like Slack's own analytics dashboard, but with a cultural/intellectual lens rather than a productivity lens. This is where the commercial ceiling is highest — B2B SaaS for culture-led companies, agencies, and strategy consultancies who care about what their teams are actually thinking about.

4. The consulting product. Run this analysis for other people. Charge for the insight, not the software. This is the lowest-investment path and the most immediate. Several people who saw the dashboard asked if I could do the same for their workspace. That's a consulting engagement, not a product, but it is revenue-generating from day one and validates demand before any product is built.

5. The data journalism angle. If the dataset were larger — even 100 people running the same analysis — the aggregate findings would be publishable. "Here is what strategists pay attention to in 2026 versus 2019." "Here is the AI cliff across 100 personal workspaces." This is not a direct monetisation path but it is an audience-building path that supports all the others.