Name: 元Facebookのデータサイエンティスト城間氏とのデータサイエンスキャリアガイド完全版 (Complete Data Science Career Guide with Former Facebook Data Scientist Joma)
Uploaded: 2021-01-14T10:23:25.000Z
Duration: 22 min 25 s
Description: VoiceTubeの動画で発音を聞きながら英語表現を覚えよう！学べる英語：

過去進行形

So a few weeks ago, I was visiting in California.

to 不定詞

When I was there, I had a chance to interview Joma Tech on just in case you don't know much about him.

過去形

Hey, went to the University of Water in Canada for a computer science degree while he was there.

Hey, did a bunch of internships in software engineering that science on P M companies like lengthening Facebook on Microsoft on After graduating, he worked as a data scientist at Buzzfeed on Facebook on.

Now he's a youtuber to now in this part of the interview that I'm going to show you now.

We basically talked about first of all, what it's like to be a data scientist and also how to become one s.

The first question I asked them was, What was he like to work as a bit of scientist at both feet and Facebook?

But aside, buzzfeed, I was basically hired to understand, like how to do how to distribute videos on Facebook.

So how it works is like, who has you know Buzzfeed is like a media company.

They make a lot of it was everywhere, and one of the platforms that they care a lot about is Facebook.

So if you've ever seen like, a tasty video of the hand and stuff, that's that's buzzfeed.

You know, they have a lot of questions, like, should we share or Sherry Cross posts, which is very like specific to like, Facebook and stuff like that.

And then also, like, what kind of videos do better than others and stuff like that.

So basically, my job is, um, is toe basically like a product analyst for Facebook videos?

What we would do is who would just look at a lot of these videos and try to understand like, Oh, why certain things happen?

Why did we get more shares unusual and then, yeah, that's pretty much it.

And then at Facebook at Facebook, I worked on like the video team, and basically what we're trying to do is we're just trying to make videos more successful on Facebook.

So it sounds like your career, like your entire career.

So, Uh, that's that's, like pretty much my choosing.

It's because even before starting at buzzfeed, I was always into, like, videos on dumb.

It's not, like, very good, but I was pretty like, I really want to be a YouTube celebrity for some reason, you know, as a very vain of myself.

But yeah, even chose, like, my career based on that.

And, you know, like I chose buzzfeed because I thought that I might be able to be in their videos in l.

And then maybe I'd get some recognition and then using that recognition, I could grow my YouTube channel.

So, yeah, I actually went to Buzzfeed for that.

Yeah, And then when I went to Facebook, my initial thought was not to be on the video team because I was like, Okay, I'm done with videos because when I switched to Facebook, I didn't want to do anything with videos anymore, But it was just so tempting because the team was like, Oh, the creator team.

I was like, Oh my God, It was like it sounds.

I'll go on that team because on Facebook, you, um you're in boot camp first, and then you get to choose what team you want to pick later.

So I didn't initially want to be in videos, but India and I still ended up there.

So you said that you wouldn't be able to share it that much stuff about the work you did at me because some of the stuff is still like, like, not public or also, like, they don't want to tell us like this strategy toe, especially YouTube and stuff like that.

But, um, yeah, I could talk about it vaguely, though.

Okay, s So are there any interesting examples of data science work?

Sometimes, I think like not every company needs a data science team.

And buzzfeed was kind of on defense of it, especially from what I was working on.

Because, like, it's kind of like like running a YouTube channel were like running a Facebook channel.

Essentially, I'm an analyst for their Facebook Channel.

And a lot of the times you can't really explain stuff with data with stats, cig, you know, like one of the questions was like, Oh, you know, should we share should be cross pulse.

Which one's better, which will get more views.

But the issue is you have to, you know, work with the people who actually managed to page make sure they post at the same time that same everything.

So basically, we did a lot of, like, crappy A B tests like it wasn't a clean, maybe test to try to discover if one thing is better than the other.

And we were cared about like a sharing better than cross posting stuff like that.

Feet was just understand the health of our Facebook pages and one of the things like it.

It doesn't even require that much data, science skills or anything.

But we just looked at, you know, how many views are we getting like?

They were getting less and less views per video, so meaning like we were making a lot more videos just to get the same amount of you as we did before.

So then we have to tell that the content team saying like Oh, by the way, this might not be working.

We need to think of different strategies to get one of you's.

So it was very similar to like what I do for my own YouTube channel.

We're just like looking at Bear's videos and various like trends and then try to capitalize on the No, no, no.

I mean, there were some other things that are more technicals, like like I automated.

They're like dashboards, automated, like e mails and stuff like that, because they were pretty behind in terms of like automation and infrastructure.

So sometimes what I would have to do is like, um, I would have to like, right slap slack bots.

For whenever something happens, I have to tell like this, like what has to tell you the content himself like that.

So I would guess that's technical because, you know, you have to call it a little bit, but, you know, like in terms of technical nous, it's not that technical, but even at Facebook, it wasn't that technical.

All I did was, like, sequel sequel queries and then use, like, tableau, toe like visualization.

And then, if I need it, like to do, like a simple inner aggression, or like, a logistic regression that I would use, like our python.

Is there anything more you might be able to share from Facebook?

Like I could say that we have a lot of data and way, have a lot of data, and it makes things a lot easier because, like, we have a lot of data engineer.

So, like, things are really, really easy to do.

All I need to do is, like, right, sequel, query.

And I could quickly, like, billions of data points just in one query and then, yeah, that's very much.

And like the job that I did at Facebook, we basically wanted to make creator successful.

So, you know, you could you could kind of, like, assume you could kind of like, um in for what I did.

But basically, we look at our platform Facebook as a whole, and then try to understand how to change the platform.

Such debt can make creators counting your creator successful in our platform.

So I guess ah, one like hypothetical example.

Might be life video versus regular video.

字幕リスト動画再生

元Facebookのデータサイエンティスト城間氏とのデータサイエンスキャリアガイド完全版 (Complete Data Science Career Guide with Former Facebook Data Scientist Joma)

stuff

strategy

basically

realize