#IDSky 🦋 Week 2 - How Does It Look?

How’s #IDSky doing on week 2? Pretty great! We’re seeing steady posts worldwide, strong US-UK participation, and even outpacing Twitter metrics. The ID community isn’t just dropping by - they’re making BlueSky their new digital home.

Week 2 Of #IDSky, How Does It Look?

Last data queried was 11/16/24 21:05 EST. From the plot, it seems like we’re still going strong! Since 11/9/24 #IDSky daily posts were maintaed at median 138 (+/-19.7). It looked like 11/15/24 was the most of the week at 164. Don’t mind the 11/17/24 since the timezone on the data was set at UTC, it’s not a complete data.

#IDSky Posts Frequency

Since I had some API access on X/Twitter left (will be expiring soon), I was able to take a look at what #IDTwitter/IDxPosts post numbers were. Here it is.

#IDTwitter/IDxPosts Posts Frequency

If there are just a bunch of numbers, they are. Basically these are total posts with #IDTwitter/IDxposts on the text. What do you think? Which is higher? #IDSky !!!

What were the conditions of query for these posts? We included retweets/reposts. So both were compared on a level playing field. But I just realized this…

RT on twitter will include all the texts of the original tweet… Repost/Quote on bluesky doesn’t, unless if the author tags idsky again. That means, the #idtwitter #idxposts counts are inflated compared to the #idsky counts on bluesky.

For example: Twitter: if this post (#idtwitter look at this article...) were retweeted twice, the counter will pick up (RT: #idtwitter look at this article...) 2 more times, making it 3 total.

Bluesky: if this post (#idsky look at this article...) were reposted twice without any tags, the counter won’t pick up the reposts, making it 1 total.

Which means… #IDsky could have a higher posting frequency than it looks!

Is #IDSky A Global Thing? Yes! 🙌

Don’t mind the color palette, or the lack of. Anything in white and without the country’s names means there were no post from author from that country. If there is 1 or more there will be color, which starts from pink, blue, purple, green. The legend on right shows the numbers of posts from the country.

Wow, this is very international! I purposefully have green to be the color of highest frequency to highlight the high posts between USA and UK.

Here is a bar plot of the top 10 countries sorted by the highest post counts. Not too surprised about USA, but it’s nice to see the USA-UK relationship is quite strong! Keep up the great bond! I don’t think this is just UK, I do believe if I were to lump the European countries, we’ll see a strong US-Europe relationship too! Not to mention Canada, Saudi Arabia and Australia too!

There were about 38 (9.7%) missing data of the country where the author’s from. There is some limitation of the method I used. Since BlueSky does not include location information like Twitter did (though these are also self-inputted), I had to resort to LLM-assisted country allocation. Since there were descriptions listed by the authors, I first used GPT4-mini to interpret the description and see if it can give me a country, if it cannot, it will automatically resort to search engine result page (SERP) API and then it will plug in the top 5 searches of the display name and description, and see if it was able to allocate a country. I will create another blog post on the details of the code if technical people are interested. I have still yet to audit this, but the USA, Canada and UK data should be quite accurate, the data on others I need to audit.

I do intend to audit the results by randomly sampling the labeeld data and assess the accuracy manually and record it to investigate the validity of the inference. From just common sense, I know Bassam Ghanem had posted quite a bit since he came to 🦋 but I don’t see his affiliation, hence I made an amendment to the plots about.

So, What Do You Think?

Well, the momentum initially was great, but to be transparent, I was concerned about the frequency of #IDsky posts initially were mainly from greetings, welcome, and introduction as people joined. However, after this week, I’m confident that the #IDSky community is here to stay! We are seeing more ID related contents, more discussions as people are comfortable with the user experience, more connectivity and visibility.

What’s Next?

We will continue to monitor the posts frequency for another 3 weeks (including this week), if the counts are the same or higher, I think we know the answer. 🦋❤️

If you like this article:

• please feel free to send me a comment or visit my other blogs
• please feel free to follow me on BlueSky, twitter, GitHub or Mastodon
• if you would like collaborate please feel free to contact me