ModScope

Calculate engagement quality based on the sub archetype (Discussion, Image, Gaming, Support, News). ModScope uses metrics such as velocity, comment depth, and even creator participation bonus.

Comment

ModScope

Inspiration

I mod a subreddit built around quizzes, and for a long time I had a problem that no existing tool solved cleanly — I needed to understand not just how many posts were coming in, but what kind. Were we getting a healthy spread of quiz categories? Were certain themes dominating the feed? Was the community drifting toward one subject area without anyone noticing?

So I built a PAW bot to answer that for my own sub. The centerpiece was a word cloud — a way to visually surface the themes and categories showing up across submitted quiz posts, at a glance, without having to manually scroll and tally. It worked. I could finally see the shape of my community's content, not just its volume.

But somewhere in the middle of building it, something became obvious: the underlying problem wasn't unique to my sub at all. Every moderator — regardless of what their community is about — shares some version of the same blind spot. Raw post counts and upvote totals tell you that things are happening. They don't tell you what's actually driving your community, which voices carry weight, whether engagement is healthy or artificially concentrated, or how your sub's character is shifting over time.

That realization is what turned a personal PAW bot into ModScope.

What it does

ModScope is a moderator analytics tool that gives Reddit communities a clearer picture of their own health and engagement. It generates configurable engagement scores tailored to how a specific community actually behaves, produces word clouds that surface the real themes and categories driving post submissions, and builds snapshots over time so moderators can spot trends before they become problems — not after.

Engagement scoring is fully preset-driven, with community type profiles for discussion-heavy subs, gaming communities, support and help forums, news and current events, and image or meme-focused spaces. Each preset reflects the actual dynamics of that community type — comment weight, upvote weight, velocity windows, depth scaling, and creator reply bonuses are all tuned to match how value is actually generated in that kind of sub.

How we built it

ModScope was rebuilt from scratch after the original PAW bot proved the concept worked. The rewrite was designed around flexibility from the start — a settings architecture that separates user preferences, calculation settings, and community presets so each layer can evolve independently. Engagement score calculations are formula-driven and transparent, so moderators can understand and trust what they're seeing rather than treating the score as a black box.

Challenges we ran into

Defining what "engagement" actually means across fundamentally different community types was the hardest problem. A support sub where OP follows up and marks a resolution is behaving very differently from a meme sub where a post hits 10k upvotes and generates twelve comments. A single formula applied uniformly to both produces meaningless numbers. Getting the preset system to reflect those real behavioral differences — rather than just surface-level weight adjustments — required significant iteration.

Velocity scoring presented its own challenge. Rewarding freshness without penalizing legitimate late engagement, and ensuring the decay curve behaves correctly across different snapshot timing scenarios, took careful tuning to get right.

Accomplishments that we're proud of

The preset system landing in a place where each community type profile is genuinely defensible — not just different numbers, but numbers grounded in how those communities actually work. The word cloud surviving the full rewrite and becoming more useful, not less, in the expanded tool. And the decision to make the engagement formula fully transparent to the moderator rather than hiding the math behind a score.

What we learned

That the instinct to generalize a personal tool is worth following, but only if you're willing to throw out the original implementation entirely. The PAW bot worked for one sub because it was built around one sub's needs. ModScope required completing refactoring all of the hardcoded logic and expand it into a fully customizable app that behaves as any sub may possibly need.

What's next for ModScope

Automated snapshot scheduling, trend visualization across snapshot history, and deeper anomaly detection — surfacing when a community's engagement pattern shifts in ways that warrant a moderator's attention. The word cloud is also a candidate for expansion into category clustering, so moderators can track not just what themes appear but how their relative weight shifts over time.

I've also begun toying with the idea of using the analytic data already captured by the backend, but introducing a user facing component that allows the end users to see certain statistics like prime posting times. That way they can take advantage of the information and interact with the sub in a more educated/inormed manner.

Built With

Share this project:

Updates