4 ms·

As a data scientist, I think most data is useless, but there is an addictive, video game-like quality to throwing lifeless spreadsheets into a machine and havin

As a data scientist, I think most data is useless, but there is an addictive, video game-like quality to throwing lifeless spreadsheets into a machine and having colorful visualizations come out. It's kind of like a very boring video game for adults that makes them feel like they're working, when they're actually just enjoying colorful abstract shapes and colors. To be honest, this is probably a sizable piece of why I'm in this line of work.

foreverCarlos 4y ago
Agreed! It all probably started with big tech promoting the “data-driven decision” paradigm. Of course, in many cases this approach is effective, but it’s not a panacea and has its limits. It’s tempting to interpret availability of any data as an amazing untapped resource, but in many cases analyzing it is just a massive waste of resources and could be more effectively replaced with traditional tools (surveys and such).
nonrandomstring 4y ago
> It's kind of like a very boring video game for adults that makes them feel like they're working, Absolutely love it. Let's distinguish a few things though. "Data science" seems like a pretty weird name. I mean, it's just "Science" right. Of course there's statistics, mathematics, signal processing, systems analysis, machine learning... all the good things that you and I are into. But how does this get huddled uncomfortably beneath the umbrella "Data science"? I think the answer is found by asking about the ends of data science, the old Cui Bono? There's the raw entertainment value you mention. It's cool to have knowledge and visualise it. Sensors, transducers, processing is fun. Then there's legibility. That is political and is about control. What most scientists are doing with data is either hypothesis testing or combing for causal relations to then abductively feed back into hypothesis generation. What most business people are trying to do is optimise, and adjust constraints and parameters. It's modelling for the most-part. It's ancient and goes back to linear analysis and regression from before the last century. Security people are looking for stress signifiers, suspicious patterns with various triggers, selectors and tripwires. Financial people want fortune telling. They want the models to extrapolate into beautiful hockey sticks. Within any organisation we may need to do one, a few, many or none at all of the above. The problem then is that "Valuable data" is such a broad, open prospect it seduces gushing, credulous administrators into valuing the process, and the tools, but not the ends.
- judge2020 4y ago
  The job description for a data scientist isn't necessarily consistent across companies, but I can imagine a good description would be to act like a real scientist in that you collaborate with product teams to figure out what experiments to run, what data to collect, and how to visualize that for both the product and business team to decide on their approach to make the company more money, ie. specializing in a/b testing and experimentation. A problem I see is creating a data science team just to follow the trends and specifically requesting they try to "study user behavior" or something in hopes the data shows some underlying trend that triggers an "eureka!" moment.
- avidphantasm 4y ago
  Agree that the term data science is strange. Data engineer seems like a better name. Science implies rigorous hypothesis testing guided by a theories of how particular systems work. Not sure that applies to most “data science” work.
  - random314 4y ago
    Data science is perfect because an actual science doesn't have the word science in it namely physics, chemistry and biology. Instead we have social science, computer science and now data science.
  - civilized 4y ago
    Already taken. Data scientists are data engineers, and data engineers are data warehouse workers and janitors.
    - nerdponx 4y ago
      Not at all. Data engineers are data engineers. Data scientists are a mix of statisticians, machine learning researchers, and data analysts.
seydor 4y ago
Sounds like finance and advertising
shagie 4y ago
> but there is an addictive, video game-like quality Have you looked at the Financial Modeling World Cup? https://www.youtube.com/channel/UCOlnCUAKLENyFC8wftR-oNw https://www.youtube.com/channel/UCOlnCUAKLENyFC8wftR-oNw Its tagline is: "Excel Esports. Yes, It's a thing"
comfypotato 4y ago
Only slightly related to your very funny comment here, you might enjoy “The Visual Display of Quantitative Information” by Edward Tufte. Not only does he help you to best make the abstract shapes and colors, but there’s an added aspect of the visuals that tickles the mind. That added aspect is how easily some information is communicated along with the image. It’s along the lines of how sometimes someone can tell a joke just by saying something as concisely as possible (even if the content isn’t humorous). Like you said, whether or not the visuals provide any value is a completely different line of discussion.
koz_ 4y ago
This definitely applies to programming as well. It's not uncommon to see people happily churning out mountains of highly redundant and repetitive code, or zealously pursuing far-reaching nit-picky refactors of dubious benefit and I think it's simply because the act of writing code and running tests and watching them go green is fun.
porknubbins 4y ago
You perfectly described why I became disillusioned with any kind of data science job. The feeling that most of the time data is used to support preexisting bias and can be arranged to say anything you want. Interesting that you know this and enjoy it despite that.
- civilized 4y ago
  Thankfully I work in an area where data science has clear fundamental value to the business.

foreverCarlos 4y ago

Agreed! It all probably started with big tech promoting the “data-driven decision” paradigm. Of course, in many cases this approach is effective, but it’s not a panacea and has its limits. It’s tempting to interpret availability of any data as an amazing untapped resource, but in many cases analyzing it is just a massive waste of resources and could be more effectively replaced with traditional tools (surveys and such).

nonrandomstring 4y ago

> It's kind of like a very boring video game for adults that makes them feel like they're working, Absolutely love it. Let's distinguish a few things though. "Data science" seems like a pretty weird name. I mean, it's just "Science" right. Of course there's statistics, mathematics, signal processing, systems analysis, machine learning... all the good things that you and I are into. But how does this get huddled uncomfortably beneath the umbrella "Data science"? I think the answer is found by asking about the ends of data science, the old Cui Bono? There's the raw entertainment value you mention. It's cool to have knowledge and visualise it. Sensors, transducers, processing is fun. Then there's legibility. That is political and is about control. What most scientists are doing with data is either hypothesis testing or combing for causal relations to then abductively feed back into hypothesis generation. What most business people are trying to do is optimise, and adjust constraints and parameters. It's modelling for the most-part. It's ancient and goes back to linear analysis and regression from before the last century. Security people are looking for stress signifiers, suspicious patterns with various triggers, selectors and tripwires. Financial people want fortune telling. They want the models to extrapolate into beautiful hockey sticks. Within any organisation we may need to do one, a few, many or none at all of the above. The problem then is that "Valuable data" is such a broad, open prospect it seduces gushing, credulous administrators into valuing the process, and the tools, but not the ends.

judge2020 4y ago
The job description for a data scientist isn't necessarily consistent across companies, but I can imagine a good description would be to act like a real scientist in that you collaborate with product teams to figure out what experiments to run, what data to collect, and how to visualize that for both the product and business team to decide on their approach to make the company more money, ie. specializing in a/b testing and experimentation. A problem I see is creating a data science team just to follow the trends and specifically requesting they try to "study user behavior" or something in hopes the data shows some underlying trend that triggers an "eureka!" moment.
avidphantasm 4y ago
Agree that the term data science is strange. Data engineer seems like a better name. Science implies rigorous hypothesis testing guided by a theories of how particular systems work. Not sure that applies to most “data science” work.
- random314 4y ago
  Data science is perfect because an actual science doesn't have the word science in it namely physics, chemistry and biology. Instead we have social science, computer science and now data science.
- civilized 4y ago
  Already taken. Data scientists are data engineers, and data engineers are data warehouse workers and janitors.
  - nerdponx 4y ago
    Not at all. Data engineers are data engineers. Data scientists are a mix of statisticians, machine learning researchers, and data analysts.

seydor 4y ago

Sounds like finance and advertising

shagie 4y ago

> but there is an addictive, video game-like quality Have you looked at the Financial Modeling World Cup? https://www.youtube.com/channel/UCOlnCUAKLENyFC8wftR-oNw https://www.youtube.com/channel/UCOlnCUAKLENyFC8wftR-oNw Its tagline is: "Excel Esports. Yes, It's a thing"

comfypotato 4y ago

Only slightly related to your very funny comment here, you might enjoy “The Visual Display of Quantitative Information” by Edward Tufte. Not only does he help you to best make the abstract shapes and colors, but there’s an added aspect of the visuals that tickles the mind. That added aspect is how easily some information is communicated along with the image. It’s along the lines of how sometimes someone can tell a joke just by saying something as concisely as possible (even if the content isn’t humorous). Like you said, whether or not the visuals provide any value is a completely different line of discussion.

koz_ 4y ago

This definitely applies to programming as well. It's not uncommon to see people happily churning out mountains of highly redundant and repetitive code, or zealously pursuing far-reaching nit-picky refactors of dubious benefit and I think it's simply because the act of writing code and running tests and watching them go green is fun.

porknubbins 4y ago

You perfectly described why I became disillusioned with any kind of data science job. The feeling that most of the time data is used to support preexisting bias and can be arranged to say anything you want. Interesting that you know this and enjoy it despite that.

civilized 4y ago
Thankfully I work in an area where data science has clear fundamental value to the business.