# Suara Restaurant Index: Sydney, October 2026

Files
- `sydney.csv`: one row per account (148 rows). Columns: id (anonymised, random order), city, followers_band
  (Under 5k, 5k to 20k, 20k to 100k, 100k+), posts_per_month, likes_per_post_per_1k_followers,
  video_share_of_posts, posts_used. No usernames or follower counts are published.
- `chart-2` to `chart-7` CSVs: the numbers behind each chart on the page. Chart 1 uses `sydney.csv`.

Method
Sample: 148 public Instagram accounts of restaurants, cafes, bars and bakeries in Sydney, found from OpenStreetMap listings, the venues' own websites and web search in October 2026. Accounts were included if they had 200 or more followers, 4 or more recent posts and at least one post in the 60 days before 14 October 2026. Excluded: hotels, directories, reviewers, creators, personal accounts, venues in the wrong city and closed venues. For each account we used its most recent posts that are not pinned and carry a timestamp, up to 12. Reference time: 14 October 2026, 06:00 UTC. Posts per month = (posts minus 1) divided by the days between the first and last post, times 30. Likes per 1,000 followers = the median likes of the account's posts that show a like count and are at least 7 days old, divided by its followers, times 1,000. The city figure is the median across accounts. Post format and caption figures are pooled across all posts used. Post hours are converted to Sydney local time (Australia/Sydney). Uncertainty: 90% bootstrap intervals (4,000 resamples of accounts). We call Sydney higher or lower than the 10-city figure only when its interval does not include that figure. Groups under 20 accounts are not shown. All figures were checked by an independent second recount. The 10-city comparison pools 1,187 accounts in Kuala Lumpur, Singapore, Jakarta, Bangkok, Sydney, Melbourne, Vancouver, Toronto, London and Dubai.

Limitations
This is a snapshot of 148 accounts in one month, not a census of Sydney restaurants. Accounts were found by search, so well-known venues are likely over-represented. We measure likes only: no comments, saves, shares, reach or sales. Likes are public counts and can include paid or bought engagement. Differences between cities smaller than the uncertainty bars may be chance. Nothing here shows that one format or posting time causes more likes.

Cite
Suara (2026). Restaurant Index: Sydney, October 2026. trysuara.com/index/sydney. Data licensed CC BY 4.0.

Licence: CC BY 4.0. Built by build.py on 2026-10-14.