Is Tableau (Data Visualization) Worth it ?

[f22raptor]

My niece is looking to get into data science. She has beginner level python skills. Would learning data visualization softwares such as Tableau accelerate her goal ? Once she is in she will still be improving her python skills.

Are Data Visualization softwares a big deal to learn ?

 

[bchip]

Yes, very big deal to know - but it wouldnt necessarily accelerate her goal.
There is a huge shortage in the market.
Data visualization helps you to see the data that already exists, they
compliment one another but its good idea to not get too confused with too many tools.

In South Africa it seems Qlik and Power BI dominate.
Tableau is big in Europe.

 

[SlinkyMike]

My niece is looking to get into data science. She has beginner level python skills. Would learning data visualization softwares such as Tableau accelerate her goal ? Once she is in she will still be improving her python skills.

Are Data Visualization softwares a big deal to learn ?

If she is proficient in Python then she will pick up day to day Tableau in a couple of hours. It's important to be able to present your findings but dataviz is a small subset of the skills required in DS.

She should focus on Python for now, lack of supply makes this a hot skill to have in DS right now and that demand isn't going anywhere. What are her stats skills like?

Is she situated in CT? I am looking for interns/early stage developers.

 

[jezzad]

We use Tableau at work for more detailed DB explanations and diplays. We do customer life time value and then segment them into specific buckets and then distribute specific comms to them based on their life time value etc..

Tableu is excellent for this as it can be embedded into a website and sent to clients for live review.

If you need to report on less complicated metrics look at Google Data Studio

Both a relatively simple to learn and get into and having them both on your CV will certainly help. There has been a big shift recently on data science and predictive data modelling.

 

[f22raptor]

Shr is based in JHB, passed maths in matric and just beginning learning python.

 

[cguy]

Shr is based in JHB, passed maths in matric and just beginning learning python.

Personally, I would consider the first steps to a career in data science being a BSC degree of some sort with at the very least 2 years of maths/app.maths and stats, preferably a degree in one of the above as well.

 

[f22raptor]

Anyone ever tried Zegami ?

Visual Data Exploration platform that enables you to explore data visually, allowing you to quickly recognize patterns, analyze information, find hidden insights and make new discoveries.
https://zegami.com/

 

[kripstoe]

I'd recommend starting with something like PowerBI. Tableau is overkill imho.

 

[f22raptor]

Isn't overkill better than underkill

 

[cguy]

Anyone ever tried Zegami ?

Visual Data Exploration platform that enables you to explore data visually, allowing you to quickly recognize patterns, analyze information, find hidden insights and make new discoveries.
https://zegami.com/

I am a little concerned that you’re putting the horse before the cart here. Every quantitative researcher or data scientist we have (and we have 100’s, including me) first obtained a solid mathematical background before doing this kind of work. This “tool chain first” approach is going to make your niece the “plot-person”, not a data scientist.

 

[kripstoe]

Isn't overkill better than underkill

Is this case, no.

 

[f22raptor]

I am a little concerned that you’re putting the horse before the cart here. Every quantitative researcher or data scientist we have (and we have 100’s, including me) first obtained a solid mathematical background before doing this kind of work. This “tool chain first” approach is going to make your niece the “plot-person”, not a data scientist.

I get you but isn't it a start to get in the industry as a "plot person" and upgrade from there ?

 

[cguy]

I get you but isn't it a start to get in the industry as a "plot person" and upgrade from there ?

It’s generally not something that is learned from something like an apprenticeship. To draw an analogy, it would be like becoming a mechanic with the hope of upgrading to an automotive engineer. Or perhaps becoming a nurse, hoping to eventually become a doctor. The actual work and fundamentals are so different, that the career change almost never happens without doing the requisite studying.

 

[SlinkyMike]

Personally, I would consider the first steps to a career in data science being a BSC degree of some sort with at the very least 2 years of maths/app.maths and stats, preferably a degree in one of the above as well.

Not sure I agree on this one. Remember that this is an industry now, it's no longer solely an academic field. What I'm getting at is that a BSc is probably less useful than solid R or Python skills given that one can work with a stats expert when necessary. It's more of a conveyor belt approach these days, and you want to be where the demand is high.

By far and away the most useful skill is data literacy and problem solving. Which algorithm for which hypothesis and what do I need to do to this dataset to extract meaningful features? That's what I look for when hiring. Not stats degrees.

I'd recommend starting with something like PowerBI. Tableau is overkill imho.

One is not inherently better than the other, there are trade offs and compromises either weay you go. Your post doesn't really make sense.

I am a little concerned that you’re putting the horse before the cart here. Every quantitative researcher or data scientist we have (and we have 100’s, including me) first obtained a solid mathematical background before doing this kind of work. This “tool chain first” approach is going to make your niece the “plot-person”, not a data scientist.

Agree. I don't know how it happened but Tableau has somehow managed to market themselves as the Data Science viz tool. Which is nonsense.

It’s generally not something that is learned from something like an apprenticeship. To draw an analogy, it would be like becoming a mechanic with the hope of upgrading to an automotive engineer. Or perhaps becoming a nurse, hoping to eventually become a doctor. The actual work and fundamentals are so different, that the career change almost never happens without doing the requisite studying.

Agree on the apprenticeship but you are arguing against yourself insisting on formal studies. An apprenticeship would replace a degree in my view.

 

[cguy]

Not sure I agree on this one. Remember that this is an industry now, it's no longer solely an academic field. What I'm getting at is that a BSc is probably less useful than solid R or Python skills given that one can work with a stats expert when necessary. It's more of a conveyor belt approach these days, and you want to be where the demand is high.

By far and away the most useful skill is data literacy and problem solving. Which algorithm for which hypothesis and what do I need to do to this dataset to extract meaningful features? That's what I look for when hiring. Not stats degrees.

One is not inherently better than the other, there are trade offs and compromises either weay you go. Your post doesn't really make sense.

Agree. I don't know how it happened but Tableau has somehow managed to market themselves as the Data Science viz tool. Which is nonsense.

Agree on the apprenticeship but you are arguing against yourself insisting on formal studies. An apprenticeship would replace a degree in my view.

Do you actually work with data scientists? What do think they do with Python or R without stats and linear algebra skills? To me, it’s like you’re calling someone a computer scientist because they know html.

 

[SkippyRamirez]

MicroStrategy might be a bit over the top for beginners, but they have a free "MicroStrategy Workstation" which is easy enough to build visualisations with.

 

[SlinkyMike]

Do you actually work with data scientists? What do think they do with Python or R without stats and linear algebra skills? To me, it’s like you’re calling some a computer scientist because they know html.

OK, if you want to have a nice discussion I'm up for that but your tone isn't currently helping to create that.

Carrying on: Yes I do. I am one. I run a DS consultancy. Have worked with various big players locally in telecoms, financial services and retail. Several successful implementations (I sound like Trump now ffs.)

I don't see that you need a BSc to gain and apply knowledge of linear algebra. I do see degreed professionals without the requisite skills to successfully design and implement Machine Learning models. That I see a lot.

 

[cguy]

OK, if you want to have a nice discussion I'm up for that but your tone isn't currently helping to create that.

Carrying on: Yes I do. I am one. I run a DS consultancy. Have worked with various big players locally in telecoms, financial services and retail. Several successful implementations (I sound like Trump now ffs.)

I don't see that you need a BSc to gain and apply knowledge of linear algebra. I do see degreed professionals without the requisite skills to successfully design and implement Machine Learning models. That I see a lot.

The fact that some people with BScs can’t do useful ML is irrelevant. A far higher percentage of those without degrees can’t either.

What percentage of your data scientists have degrees?

 

[Compton_effect]

It would be worth it - from a theoretical perspective.
Qlik is more dominant in SA, but the concepts are the same - the sooner she learns that, the better.

(Why was this thread resurrected?)

 

[SlinkyMike]

The fact that some people with BScs can’t do useful ML is irrelevant.

Fair, but then...

A far higher percentage of those without degrees can’t either.

This is even less relevant (also: since you mention a percentage - what is it and where have you derived it from?)

Your assertion is that degrees are needed for Data Scientists.

I'm saying that they prove nothing. Problem solving skills (creativity is severely lacking in rote learners IMO) and coding skills are far more valuable in my experience (which is 'extensive' given that this is an emergent field.)

You need the understanding of the principles and the ability to implement them, a BSc does not equate to this. The ability to do this is mutually exclusive from any degree. That is my assertion.

Out of interest: how long have you been in out of university and in the workforce?

What percentage of your data scientists have degrees?

0.