New Insight: Who was your first Twitter follower/First Facebook friend?

[anildash]

One-liner

Who was the first person that followed you or befriended you on this network?

Full explainer

We want to give you warm fuzzies recollecting whomever was the first to connect to you when you discovered the network.

Audience for the insight

Which networks: Ideally this should work for all of them, contingent on the possibility of retrieving the data, or any limits/constraints on paging through followers/friends. This would exclusively be on first-run, since it'll never change.

How often this insight runs

• [X] One-time only, if it's never appeared before
• [ ] First crawl only
• [ ] Daily
• [ ] Weekly
• [ ] Monthly
• [ ] Yearly
• [ ] Triggered by a data event

  Headline

In the beginning...

Body

• %username's first %follower on %network was %name. Time flies, huh?

  Criteria and logic

There's no baseline here, and we'll keep copy rather simple because it's only ever going to show up once.

Included elements

• [x] Headline
• [x] Body
• [X] Header image (image off to the left in side-by-side style insights)
• [ ] Hero image (giant image on top)
• [X] User(s)
• [ ] Post(s)
• [ ] Link(s)
• [ ] Line chart
• [ ] Bar chart
• [ ] Other viz
• [ ] Other graphic treatment

$my_insight->setPeople should be populated with the user that was the first follower/friend; Other than that this is really just the brief bit of copy.

[cdmoyer]

So, one issue is that we don't currently have this data. In our tu_follows table first_seen is just a timestamp for when the crawler first saw the user.

Looking at the graph explorer, twitter json and docs, neither provides a timestamp when a follow was created. The twitter docs say that normally the most recent follower is listed first, but that's not a promise and subject to eventual consistency issues and may change without warning.

So, I guess the questions are:

• Is this data available somewhere that I'm not seeing?
• Are we OK with doing this just for twitter, and accepting that it might not be 100% accurate?
• If we do this for twitter, we'll need to update the crawler and add some sort of "listing_order" or something to the tu_follows table so we can figure out which was the first user listed by the api.

[adampash]

I'll answer with what I think, though @anildash and @ginatrapani may want to weigh in as well.

Is this data available somewhere that I'm not seeing?

I don't think so.

Are we OK with doing this just for twitter, and accepting that it might not be 100% accurate?

I would say yes, I think for now it's worth giving it a try. @ginatrapani might have more considered feelings about playing in API gray area, though. Currently, docs state:

At this time, results are ordered with the most recent following first — however, this ordering is subject to unannounced change and eventual consistency issues.