Who posted 5 billion Tweets in 4 months?

happygeek 0 312 Views

The answer, it seems, is everyone. The total number of 140 character maximum messages posted to Twitter stood at 5 billion late last October (up from a billion the November before) but has leapt forward as the popularity of the micro-blogging system goes into overdrive and last night hit the magic 10 million messages milestone. Which means that in the space of just four months users of Twitter have posted a combined total of some 5 billion Tweets.

Or have they?

Well, there's no doubting the 10 billion tweets figure, that's as official as they come. However, just how that number is arrived at is open to interpretation. It would appear more likely than not that Twitter is not talking about 5 billion original Tweets, either posted as a new message or in reply to an existing one.

Instead it seems likely that those 5 billion Tweets include messages which are not original, in other words which have been 're-tweeted' by users keen to share the posting with their circle of followers on the system.

Nothing wrong with that, of course, although I would suggest that it does somewhat distort the usage statistic as most casual observers would quite rightly assume that the milestone figure relates to users sending original messages. Even if they are, also according to Twitter itself, posting some 600 Tweets per second or 50 million messages a day.

Anyone care to start betting when the 20 billionth Tweet will be posted? Oh, and if Johnny Depp will die again before that milestone is reached for good measure :)

Dani AI

Generated

Good point — big headline totals can be misleading. A single “total messages” milestone often mixes original posts, replies and re-shares, so it tells you more about activity and amplification than unique author-generated content. Re-shares (retweets and copies), automated feeds and duplicate posts can dramatically inflate a headline figure without increasing the amount of original material.

A simple, practical way to estimate how many messages are actually original is sample-based analysis:

  • collect a random sample from the period in question (representative of the public stream you care about),
  • mark and exclude obvious re-shares (metadata flags or the RT prefix),
  • canonicalize text (lowercase, expand/normalize URLs, strip tracking query strings, collapse whitespace),
  • deduplicate identical normalized text to get unique entries,
  • compute the fraction P of the sample that are unique original posts, and multiply the reported total by P to estimate originals. Also compute a confidence interval for P so you know the sample error.

Example pseudocode:

sample = collect_random_sample(N)
unique_set = set()
for t in sample:
    if is_retweet(t): continue
    text = normalize(t.text)
    unique_set.add(text)
P = len(unique_set) / len(sample)
estimated_originals = reported_total * P

Caveats: sampling bias (public sample vs firehose), bot-driven bursts, cross-posting services and evolving platform definitions can all skew results. If precise attribution matters, request the data owner’s counting methodology or a breakdown by originals, replies and retweets. That gives a defensible, reproducible answer instead of relying on a single big number.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.

Edit Preview