I'd like more people to join the broader data community
I want more people to talk to dangit!
Last week, I finally admitted to myself that I don’t have time to finish the program I wanted to write to pull and lightly reorganize a twitter accounts tweets to help me figure out if I should follow a person or not. One day I’ll get around to finishing it, because I was also in the process of writing a work journal about that process, as a p public example of my “Learning in Public” projects at work.
But admitting temporary defeat this week meant that I could check off a task that I had been putting off — going through a bunch of the accounts that followed me in the past three months and deciding if I wanted to follow them back. The list of accounts to check isn’t very long. like maybe 50ish entries, but there’s a certain amount of toil involved that just feels like it’s soooo close to being automated, but it isn’t because I’m not smart enough to figure out how yet.
Since my primary use of social media is to have fun conversations with people, I optimize for finding other people who speak and I can potentially have a conversation with one day. Roughly, this is the list of criteria by brain checks off to figure out if I want to follow:
Isn’t purely retweets - I’m not looking to overload my timeline with more links to random articles
Has some actual tweets from the person, fairly recently - it doesn’t have to be high volume, but something once in a while
Not one of those annoying “personal brand” accounts - it’s fine if they’re occasionally promoting something they’re doing, but not constantly. Again, humans, not LLCs.
Tweets need to be somewhat interesting to me (which could be a ton of things), and not stuff I find objectionable (e.g. racist content, etc, it’s a long hazy list)
Obviously, not everyone I currently follow these days fits this picture. But that’s the rough profile I’m using for new people going forward.
TIL: People who fit that profile are super rare!
At first blush, my criteria doesn’t sound THAT bad, but in practice, it doesn’t seem to let a huge amount of people through. Eyeballing the counts(because I was dumb and didn’t keep records), about ~15 new follows out of 50-60 accounts checked. That’s not too bad I guess? 30-ish percent. Surprisingly, it felt like I was looking through hundreds. That’s probably the strain of having to manually code 50 accounts straight.
I’m reminded of the 1% rule, where ~1% of an internet community are said to be the ones creating new content, with another 9% re-sharing content, and the majority of users do nothing. So hey, we’re doing better than that! Considering I’m mostly following people who have originally followed me, the chances of encountering a bot is fairly low. That should theoretically increase the signal-to-noise ratio.
But while ~30% sounds high, it certainly didn’t FEEL that way as I manually slogged through account after account, tab after tab. Every new tab held the glorious promise of a potentially new interesting human to interact with, and 2/3rds of the time it’s met with a solid wall of disappointment. =(
Many are just people who are read-only, so they have practically no tweets within the last year, or ever. Then there’s another group that’s all retweets without any comments. It’s just a continuous feed of articles. I assume they’re curated to some extent, but wasn’t interested in devoting time to find out.
There’s also a small group that’ll say things, but are very self-promote-y, or somehow affiliated with a workplace or organization and thus are very curated and “on-brand”. You almost hear them speak in a robotic corporate tone and it’s not particularly inviting engagement.
All of those are using Twitter to fulfill some sort of need in their life, and I’m not saying it’s bad to be any of those types. It’s just not my cup of tea.
The one big thing that I realized while sifting through all of these accounts was how I wished there were more interesting accounts to interact with. I also realized that while I often recommend that people new to the data science field to participate in the data community on Twitter, I don’t really explain what I mean when I say that. I’m sure a bunch of people were quire reasonably thinking “well that’s easy to say, but HOW?” It’s like all those pesky career counselors back in my college days were saying “you gotta network!”to the blank stares of the student body.
So in a completely self-interested attempt to find more fun people to talk to on Twitter, I’ll try to correct this oversight.
From here on out, I’m primarily speaking out of my own experiences, which obviously are not yours and your mileage may vary.
What do I mean by “participate in the data community”?
I think if I were to sum up a good optimization function for participation in a social network, it’s this — have as many meaningful interactions as possible. You get to largely determine what counts as “meaningful” in this context.
Since I’m talking about a social network, interaction is essentially required. I’m counting read-only lurking as a form of interaction, albeit one-sided. It’s not helpful, nor useful, to be on any social network and have no interaction at all. Let’s save the solipsism for philosophy class.
The main decision point is more about where to define what “meaningful” actually is. It could be reading, or sharing/reposting material, commenting, having conversations, and even posting original content.
I think part of the journey of finding your personal public voice on the internet is figuring out exactly where on the spectrum of activity you want to be. Over time that comfort level will also change. At first you might just be comfortable with asking questions in an existing conversation. Later you might start answering questions, or sharing and commenting on things. Finally you may start making your own meme videos.
In my case, I tend to focus on having interesting replies and conversations with other people, so I tend to reply more than I send open tweets out. But I also balance in posting my own content as a way to invite conversations and sharing things out on rare occasions because I think something is cool.
What about new/original content?
By “original content” here, I mean anything that isn’t purely sharing something else — the contents come from your brain and into the machine. It doesn’t have to be some epic new crystalization of pure creative genius, it could be something as pedestrian as “Good morning”. So long as you made the effort of thinking up and inputting it. No botting. Also please don’t get ideas about using GPT3…
There’s plenty of different ways to think about posting original content. I’m a master of none of them, I’m a small timer. If you can crate new memes, videos, or post technical content, go learn from the skilled data memers out there. What they do is beyond me.
My primary usage of postings messages comes from my many long years of using older shouting-into-the-void techology: IRC. Younger generations may recognize it as an open standard Slack and Discord before chat got shoved into proprietary browser apps running on your desktop just so you can send gifs to destroy each other’s productivity…
The communities I used to hang out on and eventually helped manage were small-ish, on the order of a hundred or so people, with a small group of regulars spread across the globe. There wasn’t enough people for constant activity, and most of us also left our computers on continuously so you’re never sure if there were other people actually awake amongst a sea of lurkers. How did we manage to build a community and have conversations in such an environment?
The most resilient method I’ve observed was people would send out a kind of signalling message, a trial balloon if you will. The information content of those messages was usually not very important, but the pragmatic use of it was to signal to everyone else who may have the window open on the side that there was someone present. It’s an open invitation to chat to anyone who happened to be looking at the IRC client notifications. Sound familiar? I’ve been in Slacks and Discords that still have this dynamic. Twitter follows this to a lesser degree too.
Surprisingly, this low-tech solution worked quite often when used during hours of the day that other people happened to be on. If someone, anyone, responds, suddenly a conversation has started. Then more people are likely to see that there’s some activity, which increases the chances that more people will join in because there’s more content to latch onto. Before you know it, a lively organic conversation has taken root.
This whole group dynamic, which isn’t particularly novel or surprising phenomenon, reminds me of this ancient short TED talk (yeah, yeah I know) about creating movements.
Was there any patterns to the content of these initiating messages?
Some… but it varied with the group.
The general idea was to post stuff that people are likely to react positively to. So “here’s a cool relevant thing I found” worked well, and takes little effort. The established regulars could get away with a simple greeting. In one group, we literally adopted the phrase “Ruuuuuu” as a greeting in running gag that lasted about 15 years… Newcomers might start with questions, or say “Hi!”. It doesn’t take much.
Things that were hard to respond to, like random quotes from some obscure book, things that come off as inauthentic posing (“Here’s me making a deep pithy statement” type posts, you know what I mean), a rantabout something no one else cared about, spam, trolling, all those generally didn’t work out very well. Even if it sparked activity, it was mostly backlash from the community at large, which is attention you don’t want.
So the practical move is to just occasionally send out messages that may potentially be interesting and see what sticks. Put a bit of your humanity into it, since shared excitement or common experience is always a potential bonding point even with strangers. Over time, as your audience grows and responds more, you’ll get better at it.
But hey, I don’t have an audience! I’m new at this!
Yup, but in a sea of bots and lurkers, you need to signal that you’re a human that wants to engage in conversation somehow.
Early on it’s useful to join in on the conversations of other people, since an intelligent or witty reply is also a clear sign of human-ness. But at some point you should transition to starting some conversations yourself. Plus, if you don’t practice while there are few eyes on you, you’re not going to get into the habit of being mindful of what you do post when you make it big.
OK, but what about content?
I dunno. That’s up to you. What do you care to have conversations about?
I mostly do cooking, tech, data, games, translation, what the 1yo is learning to do, how impossibly humid it is outside right now, and a bunch of other weird hobbies and interests. That chaos works for me and likely not for you.
More telling are the things I generally avoid wading into. I’m not in the mood to argue about politics on the internet, so I generally keep my opinions to myself except on rare occasions. I avoid directly speaking about work to keep from getting into trouble by accident. No kid or family pics, no sketchy material, etc.
Luckily, I’m not in a position where the effects of my opinions and views may have an effect on others. For example, I have friends that run small companies that have to be careful about what they post on social media, because a public backlash to a comment in a polarizing issue from their customer base can cause harm to the livelihoods of employees. We can’t all be Elon Musk and tweet whatever controversial thing we want and get away with it.
Just as you have a choice of what to post, you can pick your battles over what not to say. Oftentimes that's the more important decision to make.
Keep your conversations public
Starting out, you might be tempted to ask a question in private. One common example is “can you help tell me more about your job?”. I'd say that unless you're sharing information that you don't want public, it's pretty much always better to ask such questions where other people can see. Because other people can join in, others who don’t dare to ask questions can benefit too. That's a good thing.
It also can be more rewarding to have conversations with less famous and followed people. Someone with thousands of notifications to sift through can easily skip you. But there are plenty of great, smart people who have much smaller follower counts that probably won't.
Generosity pays off
One of the defining features of the data community is the free flow of knowledge and support. There's no advantage to hoarding this stuff for yourself. The best thing to do is contribute and give this stuff out in some way, even if its just encouagement or empathy, because eventually it does come back around.
Working at small tech startups for over a decade, I've been hit with the layoff axe on multiple times. Two of my previous jobs essentially came through the help data community. Back then I was just some guy occasionally chatting about making shoddy excel charts, cameras, and go. But those random tenuous connections and friendships came through in surprising fashion. You never know where things will lead so generosity helps that a ton.
So please, go make some new data friends out there, and have fun and be safe while doing it.
Nice advice, it's hard to start in twitter, maybe there is so much noise that everything we think to post just never seems right. @cuasiperfecto.
The Twitter data community seems generally fantastic, but it takes so much work to avoid the negative and toxic parts of the platform. I want to spend hours immersed in data and visualization and stats, as told by the smartest, most thoughtful people. But that's a challenge when the politics and the hot takes pop up everywhere, like attentional black holes.
Do you have any automatic-ish ways of addressing that? I wish there was an NLP uBlock Origin that just filtered out controversy!
Thanks for a great article. Recent subscriber, love your blog!