How to Read dstat Metrics

This page explains, in plain language, how dstat calculates the public metrics shown on the site and in reports: which period is used, how the core formulas work, and how to compare metrics correctly.

The numbers in the examples below are illustrative. They are meant to show the mechanics of the calculation, not real brand values.

Below you will find separate pages for every supported platform. Each one shows formulas, examples, and reading notes for that specific source.

dstat first determines the selected analysis period. It then counts how many publications fall into that interval and divides that count by the number of days.

period_days = number of days in the selected period

ppd = posts_count / period_days

• *_ppd shows the average number of publications per day within the selected period.
• The same idea is used for separate content streams such as vk_video_ppd and vk_clips_ppd.
• If there are no publications in the period, the metric does not turn into an invented value.

For social and video platforms, dstat first sums likes, comments, shares, and views for the period, then builds percentage metrics either from views or from audience size.

actions_sum = likes_sum + comments_sum + shares_sum
views_sum = total views
followers = audience size

like_rate_by_views = likes_sum / views_sum * 100
interaction_rate_by_views = actions_sum / views_sum * 100
like_rate_by_followers = likes_sum / followers * 100
interaction_rate_by_followers = actions_sum / followers * 100

• interaction means the sum of likes, comments, and shares.
• share combines reposts and forwards when the platform supports both signals.
• If views or audience size are unavailable, that percentage metric stays blank.

Not every metric is calculated the same way. Some metrics show total volume for the period, while others show what a typical publication looks like.

median = the middle value of a sorted set

• A median is useful when one viral post would otherwise distort the average too much.
• Totals answer “how much in total,” while medians answer “what is typical for one publication.”

• Compare the same platforms and the same denominator types: by_views with by_views, by_followers with by_followers.
• Do not compare app-store ratings to engagement metrics from social platforms.
• A blank value usually means there was not enough source data for the calculation, not “zero performance.”
• If you need detail, open the platform subpages: they contain worked examples for each source.

• *_ppd is about publishing frequency.
• by_views and by_followers should always be read literally from the name.
• Totals show overall volume, while medians show the typical publication.