How BellyUp! ATL Works

BellyUp! runs on four steps, in order, and every page on this site is an output of one of them.

1. Collect

We continuously gather publicly available and directly measurable information about Atlanta venues: restaurant websites, published menus, structured data, hours and location information, public events and specials, public social activity, aggregate reputation signals, and technical accessibility to search engines and AI crawlers. Once a venue claims its profile, whatever the restaurant tells us about itself joins that set.

Website checks are run by our own scanner against each venue's own site and its robots.txt. It identifies itself honestly as BellyUpDiscoveryBot, requests at normal rates, does not log in, and does not touch anything behind a form. It stores the pass, fail, or unknown result of each check and nothing else. If you run a venue and would rather it did not run against your site, tell us and it will not.

2. Structure

Raw information gets translated into discovery categories. Instead of only knowing that a restaurant exists, we build an understanding of what that restaurant could reasonably be discovered for. Some of that resolves automatically from a venue's attributes. The rest ... an origin story, a service ritual, a historic room ... is tagged by hand against something we can point at.

The category taxonomy is public and machine-readable at /data/discovery-categories.json, and the venue dataset is public at /api/places.json.

3. Measure

Then we check whether any of it is actually visible. Which venues are discoverable for a given category. Which experiences are poorly represented online. Whether a machine can understand a venue's specialties. Whether its information is consistent and reachable. Which important categories are missing from the city entirely.

That measurement is published as the Atlanta Discovery Report and as a discovery profile on every venue page. Nobody can buy a position in either one.

4. Publish

The same dataset becomes the reading experience: discovery category pages, neighborhood pages, guides, and the audience sites below. One collection system feeding many surfaces is the whole design, which is why adding a new lens does not create a new editorial workload.

One venue, one record

Every restaurant, bar, and nightlife spot in the network lives as a single record on bellyupatl.com, tagged with the audience, dietary, experience, and discovery attributes that describe it. There is no duplicate content across sites. Every category subdomain (Vegan ATL, BBQ ATL, Bars ATL, Family Friendly ATL, Nightlife ATL, and more as they launch) reads that same data and filters it for its audience.

A venue can legitimately appear on more than one of them. A BBQ joint with a serious cocktail program belongs on both BBQ ATL and Bars ATL. That is by design, not a duplicate.

Community submissions

Anyone can submit a spot. Submissions are reviewed before they go live, and once approved they are always attributed to the person who submitted them, as a community recommendation and never as a BellyUp! ATL claim. They add to the picture. They are not what the site runs on.

Verification and honesty

Every venue record notes when it was last verified. Hours, menus, and ownership change, so confirm directly with the venue before you go. A detail we cannot verify is left empty rather than guessed at, and it then shows up in that venue's discovery profile as a gap. If sponsored placement ever appears on BellyUp!, it will be labeled as sponsored and kept out of every measurement we publish.