#SaveWarriorNun News & Data 👼 2023-03-18
How do we measure tweets?
Watch this issue on YouTube:
There have been a few questions lately about tweet counts, so let's address those questions. First, how do companies get Twitter data? There are two fundamental ways: first, you get data from Twitter itself via the Gnip API (application programming interface). Gnip is sold to enterprise software companies; it's reassuringly expensive, on the order of tens of thousands of dollars per month paid by each company to Twitter. In exchange, these software companies get almost unlimited access to Twitter's databases and near-realtime snapshots of the data as Twitter has it.
The second way is to collect data via Twitter's regular API. This is far more cost effective, but it comes with some tradeoffs. First, when you collect data from the regular API, you have to collect at a very high frequency and you have to store the data yourself. When I first started working on the campaign, I set up a server and had it collect data from the Twitter API every 5 minutes. Even with that advanced setup, it only collected about 55% of the tweets available, especially during peak times, because the API only returns a certain number of results at a time.
This means that if you use the non-enterprise API, you sometimes have to go back later and try to find the missing data - and sometimes that results in tweets being counted twice.
In the campaign, you'll hear about two major tools being used. This newsletter uses data from Talkwalker.com, which in turn uses the Gnip enterprise API. As a result, each day when I send out the newsletter, I'm getting a live look at the #SaveWarriorNun hashtag as it stands, present and past. Which means that if a tweet is deleted, or an account is suspended, those tweets also disappear. Because Talkwalker.com uses the enterprise API, it's most in sync with what Twitter itself sees.
The second tool you'll hear about a lot is Tweetbinder. This is a small software company based in Spain that I believe - because I don't know, I've never talked to anyone who works there - uses the second method to gather its data. Folks using it for the #SaveWarriorNun campaign have to do a LOT of manual work to manage Tweetbinder data, recording their observations in spreadsheets, etc. This introduces a lot of potential for inaccuracies.
So the question is, which is the better measure of the campaign? If #SaveWarriorNun were a paying client, I would advise them to use the Talkwalker data because it's the most accurate. However, a paying client would also probably be using the data in more ways than just keeping track of the overall campaign, so there would be higher levels of risk for inaccurate data. Because #SaveWarriorNun isn't a paying client, isn't a centralized organization, and isn't doing anything with the data other than measuring the general progress of the campaign, the risk to the campaign of inaccurate data is relatively low. Would it be nice if all the tools, software, and people keeping track of the data were in agreement as to what the totals for any given period of time are? Sure. But is there risk of substantial harm if the totals aren't in agreement? No. Nothing bad happens if one person says it's X tweets, another person says it's Y tweets, and a third person says it's Z tweets. A streaming service isn't going to make a decision to renew our show solely on the number of tweets we generate.
The key takeaway is that directionally, all the folks contributing their voluntary efforts to measuring the campaign are measuring in the same general direction - up. When it comes to expending energy and effort for measurement, we should be focused on creating more tweets and more social media content on every channel that will have us. That's where our energy and effort is needed most, to ensure that folks know we're not going away.
Christopher Penn, Chief Data Scientist, TrustInsights.ai and head WN Substack writer
Please share and recommend this newsletter to fellow Halo Bearers!
Today's Data
All data below is as of 23:00 UTC-0.
HEY THERE LORENA is today’s phrase - pitch in!
Yesterday’s off-day still came in around 31K.
We’re doing great with coverage this week! Keep on clicking on articles.
Your Attention Please
A reminder: One of the major aims of the campaign is to get more mainstream media coverage. Mainstream media publishers don’t care about retweets - they make their money on page views and even compensate writers based on how many people read their articles. The more we help them, the more they help cover us.
Please click on the link to these news articles and scroll to the bottom of each article. You don’t have to read it or hang out after you’ve scrolled to the bottom unless you want to.
👉 Fans Accuse Netflix, Disney+ of Targeting LGBTQ+ Positive Shows (www.cbr.com)
👉 ‘You’ Dethroned Again In Netflix’s Top 10 List By A Surprising New Show (www.forbes.com)
👉 12 Best Fantasy TV Shows on Netflix (www.pastemagazine.com)
Around the Fandom
Space update tomorrow:
Today's phrase:
Vote for William Miller's (Adriel) band:
Episode 3 of Out of Habit is available:
An update on Had:
Signal Boost
Some top posts that you might want to check out and reshare.
