Author Topic: part-time statistics or data analysis gigs - need career advice  (Read 2008 times)

milliemchi

  • Bristles
  • ***
  • Posts: 297
I want to be able to switch to part-time work at some point by taking on projects and working part time and/or on and off. Data processing, analysis, knowledge synthesis (data mining+), and creative ways to present data in order to generate or support hypotheses are my strong suits. Also, I find myself policing the statistics on the majority of papers I review, so I figure those are skills in short supply, and I am ahead of the crowd. My idea was to get an online MS in applied statistics, for credibility in recessions, etc. I learn best when left alone with a book, so the degree itself ($$) would really only be for show - very expensive and potentially useless unemployment insurance.

However, I have recently poked around various statistical packages and I found that the main packages cost big money and crunch big data. I figure the companies that can afford the software are likely to already have in-house statisticians, because statistics is often the business model, not means to an end. Now, I am questioning whether paying or even studying for the degree makes sense if it possibly does not offer required skills. Much more importantly, I am questioning the whole part-time plan. In my old vision, there were data processing gigs that I could get here and there. But there possibly could be few side projects for someone without prior big data experience.

1. True/untrue? Thoughts? Are small data projects still available somewhere?

The online program at Penn State (well ranked, ~$25-30K) offers the following elective courses:

Design of Experiments
Epidemiology Research Methods
Design and Analysis of Clinical Trials
Statistical Analysis of Genomics Data
(possibly useless as clinical projects have professional (PhD) statisticians in-house)

Applied Time Series Analysis (for e.g., finance)
(likely useless as market modeling has been using partial differential equations 20 years ago, now they have likely moved on to even more esoteric methods)

Applied Data Mining and Statistical Learning (Data Mining tools: exploring data with regression, PCA, discriminate analysis, cluster analysis, classification and regression trees (CART).)
(possibly useful on big or small data)


2. Is there a market for part-time projects with any of these skills? If not, is there something else that I should be focusing on in terms of statistical skills?

3. Would the last course give me enough skills to get a credible start on big data processing?

4. Is it possible to contribute meaningfully with just one course in time series analysis? Finance I figure not, but are there other uses?

5. Does anybody have experience with the freelance market for statisticians? How workable is the part-time plan at all, data big or small?

6. Given that I have a quantitative STEM PhD from a very highly ranked university, and will have 20+ years of experience in quantitative science research, can I put the degree on my CV/LinkedIn as 'self-study' and sound credible? Is there a way that this is a good idea besides saving money?

7. Could my data mining/presentation skills, together with some non-basic statistical skills (such as the last course) be useful somewhere (e.g., presenting business plans for small businesses, helping smaller research projects in non-STEM fileds)? Could I just skip the whole online statistics program and save myself time and money? if so, what other skills would be useful to acquire? Which statistical skills would be central?

I'd be thankful for any insight.
« Last Edit: August 01, 2017, 11:04:41 PM by milliemchi »

maizeman

  • Handlebar Stache
  • *****
  • Posts: 1623
  • Location: The World of Tomorrow
Re: part-time statistics or data analysis gigs - need career advice
« Reply #1 on: August 01, 2017, 11:10:05 PM »
A very interesting set of questions! I wish I had more advice to offer but hopefully some other forum members will have relevant experience to share.

I have recently poked around various statistical packages and I found that the main packages cost big money and crunch big data.

What software packages are you looking into? Most of the complex stats I do (or get actual statisticians I collaborate with to work on) gets done in R using a bunch of open source packages.

Also, maybe building up an interesting portfolio of data mining/visualization projects using websites like Kaggle would help you build your credibility to potential employers without spending the time and money on an online MS.
"Itís a selective retirement," Richard explained, "a retirement from boring s**t."

My source code & my journal

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #2 on: August 01, 2017, 11:24:46 PM »
IIRC, I was looking at SAS, SPSS, and some other ones. A lot of cheaper/free packages are specific to what the author (often a university group) needed.

Now I see that the most expensive STATA is only $2-3K for a permanent license, so that could be a good option for freelancing. That one didn't come up in my original search a few months ago. Can anybody give a quick review on STATA?

ETA: Oops, STATA is another $1000+/year to renew.  That would require 15-40h/year just to cover cost (being taxed at ~50% in my current bracket). Not bad, but not great either.
« Last Edit: August 01, 2017, 11:45:10 PM by milliemchi »

FactorsOf2

  • 5 O'Clock Shadow
  • *
  • Posts: 14
Re: part-time statistics or data analysis gigs - need career advice
« Reply #3 on: August 02, 2017, 07:33:08 AM »
Hi milliemchi!

My feeling is that for people breaking into data science / data analysis it is strongly recommended to pick up R or Python over some of the proprietary packages you mentioned. Of those two, R is pretty steadily losing market share to Python in data science and machine learning realms (Python is a better end-to-end programming language and IMO is much more intuitive). However my husband is a statistician so I can confirm that in academic statistics R is more heavily used.

With a STEM PhD and a relevant portfolio I absolutely think you can skip the degree program. It seems like the quantitative analysis world is filled with people who made the transition from STEM academia without doing another degree. Hop over to upwork.com and search for jobs in the areas you are interested in. They have a very significant data analysis / data science / machine learning presence and taking a few smaller gigs on there is a great way to start building your portfolio without having to work for free. Some of those gigs work with big data and the associated ecosystem but most do not.

kwarden13

  • Stubble
  • **
  • Posts: 185
Re: part-time statistics or data analysis gigs - need career advice
« Reply #4 on: August 02, 2017, 07:47:00 AM »
Most companies using SAS, have Business Intelligence departments. I have worked at a few Fortune companies in this area. There was already a statistics dept, business intelligence dept, decision science dept, etc.. You would be better off maybe looking into smaller boutique consulting firms to see if you can do something part time.

Also, I would not bother with SPSS. SAS and Python are the two used frequently in my experience. Also, learning more dashboarding type tools is good so you can present your data/findings. Tableau, JMP, etc..

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #5 on: August 02, 2017, 09:57:17 AM »
Hop over to upwork.com and search for jobs in the areas you are interested in. They have a very significant data analysis / data science / machine learning presence and taking a few smaller gigs on there is a great way to start building your portfolio without having to work for free. Some of those gigs work with big data and the associated ecosystem but most do not.

Whoa! I searched upwork.com job postings, and this is what I find: D3.js, Tableau, main KPIs, Power BI, kibana, elasticsearch... Definitely not my skill set.  I don't even know what any of these are. This is just from a quick glance at the first page of /data-science-analytics/sc/data-visualization/.  data-science-analytics/sc/data-mining-management/ is a little better, but there is a lot of web scraping work - blah.  data-science-analytics/sc/quantitative-analysis/ is more promising. However, the budgets are low, it makes me think that the posters are trying to recruit from outside the US to get cheaper labor. I would likely be working for free or just about, which could be OK if I'm building a portfolio.

I'm sure there are other labor markets that pay regular wages. Does anyone have experience actually getting paid work, not portfolio building?

dinkhelpneeded

  • 5 O'Clock Shadow
  • *
  • Posts: 74
Re: part-time statistics or data analysis gigs - need career advice
« Reply #6 on: August 02, 2017, 11:45:24 AM »
Hey - I'm a statistician so I could answer this!

If you have a STEM PhD you absolutely dont need a masters, infact people might question it. If I were you I'd position it as - statistics piqued your interest during the PhD program and then you picked up most of it yourself. I would first decide what kind of satistics datascience you want to do. There are two schools of applicability - there's the traditional banking, insurance side  (SAS, R) of things where traditional statistics takes precedence, then there is machine learning/AI (Python, D3) side of things at technology companies where there's a mix of computer science with statistics. You need to learn to program regardless.

If you're leaning toward the machine learning side then start participating in Kaggle competitions build up your brand and resume a bit.
Regarding part time work - i can confirm what you are finding its almost non-existent. You will find report monkey type roles but not true data science roles. You will find one-time work in upwork but it likely will not make enough money to sustain a regular paycheck. I am a new mom with < 1 year old baby so I did consider going part time on gigs but i couldnt make it work.
I work in the banking side of things (6 years) and have a masters in applied statistics, I recently switched to a role in machine learning, so if you need specific advice let me know.

Jouer

  • Stubble
  • **
  • Posts: 220
Re: part-time statistics or data analysis gigs - need career advice
« Reply #7 on: August 02, 2017, 12:26:30 PM »
Another statistician here. Absolutely no need for further degree if just for paper.

I've done plenty of freelance work developing statistical models for companies but it's usually through small consulting companies that either do not have the resources in-house or their resources are overwhelmed. I got all of those gigs from networking - usually word of mouth from former colleagues or clients.

Agree with others - SAS/SPSS et al are losing market share to R/Python et al. There are tons of free online courses and other resources for both R and Python - just roll your sleeves up and learn them. Any analyst worth their salt can learn software on their own, since they will be the naturally curious type.

Power BI and Tableau are data visualization software. What are you using for data viz? Tableau has public software online so you can teach yourself. Many youtube videos online to help - that's how I learned. You'll need one of these and/or be excellent at predictive modeling / machine learning.

Software is software. You can learn any of them. Your analytical skills are what's important.

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #8 on: August 02, 2017, 04:07:00 PM »
dinkhelpneeded, Jouer, this is extra helpful.

Quote
If you have a STEM PhD you absolutely don't need a masters, in fact people might question it.

Can you elaborate on this? I would like to get a very good understanding of this before I can nix the MAS idea without later self-doubt and reexamination.

Quote
I work in the banking side of things (6 years) and have a masters in applied statistics, I recently switched to a role in machine learning, so if you need specific advice let me know.

Was the MAS useful for the skills of for the paper? Most of the non-electives seem to be relatively basic stuff (at least in content), which makes me think the program is designed to get people going who don't have prior statistics knowledge. Is this a fair impression? That might explain why it's not useful to STEM PhDs.

Quote
Agree with others - SAS/SPSS et al are losing market share to R/Python et al. There are tons of free online courses and other resources for both R and Python - just roll your sleeves up and learn them.

Are there extensive, useful, compact (not getting 10% of functions from this library, and 20% from another, etc.) statistical libraries for these languages? I would expect that for R, but am not familiar with what Python has to offer.

Quote
Power BI and Tableau are data visualization software. What are you using for data viz?

Right now, I am using IDL for all data analysis. Despite self-promotion, I find IDL to have unfriendly visualization tools. From documentation/help I see that they have made changes to that set over the past 5 years or so, but frankly it is not important enough for my work now to bother exploring that - especially since I won't have the license if I leave the lab, and few other places use it. Plots for papers/grants are made either in IDL or, if simple, Excel. My skill is not in graphic design, but in picking what to present and how, what to correlate with what, what proves/disproves a point, what is relevant to problem at hand, etc.

I checked out Tableau, and I checked out Power BI. While Tableau is on a monthly racket payment plan like so much software, Power BI Desktop is free. Is this good, or too good to be true? If it's a good software package, I would start there, as it would help my research too.

Quote
Kaggle competitions

Looks like fun! It seems like it could be used to learn techniques, I guess that is the point, yes?

dreams_and_discoveries

  • Pencil Stache
  • ****
  • Posts: 793
  • Location: London, UK
Re: part-time statistics or data analysis gigs - need career advice
« Reply #9 on: August 03, 2017, 01:48:41 AM »
I work in data, and whilst there are part time permanent roles, most gigs are full time.

Anyone can manipulate data, the true value add comes from knowing what the data is and how to interpret it in line with business priorities.

I'd say small projects projects fall into two categories
(a) race to the bottom types, tiny fees and usually taken by people in LCOL countries
(b) very important projects with tight timescales, taken by people with years of data experience on CV

When I recruit I don't mind too much about degrees, it's mainly experience, aptitude and work ethic I look for.

So in short, I'd work out a way to get experience you can showcase to an employer, an extra degree does seem like an unnecessary expense.

Rachel_the_Lark

  • 5 O'Clock Shadow
  • *
  • Posts: 27
  • Location: Delaware
Re: part-time statistics or data analysis gigs - need career advice
« Reply #10 on: August 03, 2017, 04:59:46 AM »
Are there extensive, useful, compact (not getting 10% of functions from this library, and 20% from another, etc.) statistical libraries for these languages? I would expect that for R, but am not familiar with what Python has to offer.

Ehh....not really.   At least not in the level of depth I imagine you're looking for.  Different people will use different packages depending on the problem at hand and personal preference.  Majority of the trainings/introductory examples for Python I see use scikit-learn so that's a good place to start, but doesn't cover all of the ground for Python's potential.

YoungInvestor

  • Bristles
  • ***
  • Posts: 361
Re: part-time statistics or data analysis gigs - need career advice
« Reply #11 on: August 03, 2017, 05:55:57 AM »
R and Python are a given. When you are studying, you're also likely to be able to get a free SAS license for students.

I would try to pick up SQL as well, because it's fairly simple and manipulating data can be a huge part of statistical analysis.

Crazy as it sounds, Excel and VBA skills can be very well received at some smaller companies, and are very easy to learn as well.

spjulep

  • 5 O'Clock Shadow
  • *
  • Posts: 54
Re: part-time statistics or data analysis gigs - need career advice
« Reply #12 on: August 03, 2017, 06:39:45 AM »
You can teach yourself the programming side. There are free classes on Coursera (perhaps a small fee if you want the certification) that are the exact same curricula that they teach in school. For example, the Data Science Specialization from Hopkins.

Jouer

  • Stubble
  • **
  • Posts: 220
Re: part-time statistics or data analysis gigs - need career advice
« Reply #13 on: August 03, 2017, 07:23:19 AM »
Quote

Right now, I am using IDL for all data analysis. Despite self-promotion, I find IDL to have unfriendly visualization tools. From documentation/help I see that they have made changes to that set over the past 5 years or so, but frankly it is not important enough for my work now to bother exploring that - especially since I won't have the license if I leave the lab, and few other places use it. Plots for papers/grants are made either in IDL or, if simple, Excel. My skill is not in graphic design, but in picking what to present and how, what to correlate with what, what proves/disproves a point, what is relevant to problem at hand, etc.

I checked out Tableau, and I checked out Power BI. While Tableau is on a monthly racket payment plan like so much software, Power BI Desktop is free. Is this good, or too good to be true? If it's a good software package, I would start there, as it would help my research too.

Tableau Public is free. I know plenty of people making names for themselves by using the free version.

I haven't used Power BI yet. We have it at work but I use Tableau instead. To be honest, I'm more of a modeler than a data viz guy, which it sounds like you are too. Good to have the data viz stuff in your back pocket though - being able to automate the boring stuff frees up more time for the fun stuff.

Power BI is very powerful. It's part of the Microsoft suite. Oh, I second whoever said to learn some basic SQL. So you use SQL to get data from the data warehouse, run your analytical model in R, which can then be embedded into the SQL code for future use/production of the model. Then on the backend, use Power BI to automate the results.

urbanista

  • Stubble
  • **
  • Posts: 212
  • Location: Australia
Re: part-time statistics or data analysis gigs - need career advice
« Reply #14 on: August 03, 2017, 08:42:05 PM »
I highly recommend reddit data science. Tons of STEM PhDs converted to data scientists are there discussing questions you've asked.

As to part-time analytics/data science/statistics roles, they are non-existent. Partly because this type of roles require continuous upskilling/learning new things otherwise you quickly become obsolete.  I (MSc in mathematical statistics) spend 5-10 hours weekly learning new things that are not related to my immediate work. My younger colleagues spend even more time learning on the side. Some of it can be done at work, but I find myself studying 3-4 nights a week for 1-1.5 hours. You have to love your work, otherwise it's not worth it.

I found that there many in-house roles allow flexible schedule + a lot of work from home + 5 days a year study leave + learn at work.

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #15 on: August 03, 2017, 10:26:32 PM »
I highly recommend reddit data science. Tons of STEM PhDs converted to data scientists are there discussing questions you've asked.

Great, thanks.

Quote
As to part-time analytics/data science/statistics roles, they are non-existent. Partly because this type of roles require continuous upskilling/learning new things otherwise you quickly become obsolete.  I (MSc in mathematical statistics) spend 5-10 hours weekly learning new things that are not related to my immediate work. My younger colleagues spend even more time learning on the side. Some of it can be done at work, but I find myself studying 3-4 nights a week for 1-1.5 hours. You have to love your work, otherwise it's not worth it.

Ugh. I'm so glad I started this thread. No kidding, it's (future) life changing. :)

Definitely not what I was looking for. I was looking for less, not more work. Too bad. I know I would love the work, but I love my current work too. No point in switching if there is no part-time option. Given all this, it might even be much easier to transition into part-time research here where I am already, I was just looking to do something I haven't already spent 20+ years doing. It gets boring at some point.

Alps

  • 5 O'Clock Shadow
  • *
  • Posts: 22
Re: part-time statistics or data analysis gigs - need career advice
« Reply #16 on: August 09, 2017, 04:20:25 AM »
Quote
As to part-time analytics/data science/statistics roles, they are non-existent. Partly because this type of roles require continuous upskilling/learning new things otherwise you quickly become obsolete.  I (MSc in mathematical statistics) spend 5-10 hours weekly learning new things that are not related to my immediate work. My younger colleagues spend even more time learning on the side. Some of it can be done at work, but I find myself studying 3-4 nights a week for 1-1.5 hours. You have to love your work, otherwise it's not worth it.

Ugh. I'm so glad I started this thread. No kidding, it's (future) life changing. :)

Definitely not what I was looking for. I was looking for less, not more work. Too bad. I know I would love the work, but I love my current work too. No point in switching if there is no part-time option. Given all this, it might even be much easier to transition into part-time research here where I am already, I was just looking to do something I haven't already spent 20+ years doing. It gets boring at some point.

Hahaha that's funny. You should know that n=1 is not a significant sample size! (I don't want to be mean, but I honestly laughed out loud when I read your answer...)

Anyway, I also made that transition and while there is definitely constant learning involved, I never do that in my free time (and I never work overtime). It probably depends on your environment. If you work in a team you have other people you can ask. Just write their answers down and reuse them, and you will be ahead of the curve. It is probably a lot harder to get good experience if you just do small projects by yourself. There is also a lot of data engineering as part of data science, and I'm happy to be part of a team where others do most of that. You sound like you might also not quite enjoy that part, if you would even have to learn programming first (sometimes the hardest/most time consuming part is just to get the data to load).

However I do work 80% hours. At least here in Europe it is definitely possible to downsize to 60-80% as a data scientist.

I would add the disclaimer though that many, many projects don't actually involve complicated statistics or procedures. These are just too time-consuming to implement properly (i.e. so they're robust in real-life situations) that it is most often not worth it from a return-of-investment point of view. Also, from what I've seen many big non-tech companies do have a lot of data, but the most they are doing is a trend analysis. Lots of low-hanging, boring fruit still around. As a poster above said, knowledge of the actual statistics is not that useful. Most people working in this field know enough to be able to read and understand the math if they need something specific. It's the real-life experience that counts, about what is viable to implement and what is actually useful to a business.

trollwithamustache

  • Bristles
  • ***
  • Posts: 315
Re: part-time statistics or data analysis gigs - need career advice
« Reply #17 on: August 09, 2017, 09:11:44 AM »
There is a lot of talk about some high end statisticery here. I would be grateful if you could just keep some of the yay-whoos I work with out of trouble  with the stuff excel does for them and why a couple data points with a "good enough" R-squared value may not justify millions of dollars in spending on a decision.

On the corporate side, this 6 Sigma and now lean 6 Sigma produces a lot of need for statistics and many fall short on semi correctly handling their data. If you have academic level skills you are certainly qualified for this.

Goldielocks

  • Magnum Stache
  • ******
  • Posts: 4957
  • Location: BC
Re: part-time statistics or data analysis gigs - need career advice
« Reply #18 on: August 09, 2017, 09:25:28 AM »
Can you piece together full time work, with long gaps between employers?  e.g., 18 months working, 1 year gap, that sort of thing?

I am not sure how projectized your area of data science is, but the marketing world Big data definitely is projectized (business cycles with whims and new ideas to implement in Marketing), so that although many people do stay on year after year (or get hired away with that intent), it is quite ordinary to quit as a project winds down, then reappear with another project or employer a year or so later.

dinkhelpneeded

  • 5 O'Clock Shadow
  • *
  • Posts: 74
Re: part-time statistics or data analysis gigs - need career advice
« Reply #19 on: August 09, 2017, 01:26:30 PM »
dinkhelpneeded, Jouer, this is extra helpful.

Quote
If you have a STEM PhD you absolutely don't need a masters, in fact people might question it.

Can you elaborate on this? I would like to get a very good understanding of this before I can nix the MAS idea without later self-doubt and reexamination.

Quote

I meant to say during your interview you are going to be asked why you decided to do a masters after your PhD and how convincing you are in your story telling will determine how much they buy your argument. A PhD is considered final and you acquire all the skills you would need to do quantitative research so going back for a masters unless you have a specific goal/story is moot. A lot is going to hinge on that one answer, its tone, its content etc. too much for me personally. I would rather the interview be more focussed on gauging my skills and more positive tone. I interview for my company and in the end the "gut" feeling they walk away with is what matters - along with technical skills ofcourse. So if you are in general a really good sauve communicator I think you'll be ok, if you tend to be even a little bit nervous then the battle will be lost.

I work in the banking side of things (6 years) and have a masters in applied statistics, I recently switched to a role in machine learning, so if you need specific advice let me know.

Was the MAS useful for the skills of for the paper? Most of the non-electives seem to be relatively basic stuff (at least in content), which makes me think the program is designed to get people going who don't have prior statistics knowledge. Is this a fair impression? That might explain why it's not useful to STEM PhDs.


Yes the STEM PhD should've already taught you how to do quantitative research. Masters starts with knowledge of basic probability theory and algebra, nothing much else is really a prerequisite.

Quote
Agree with others - SAS/SPSS et al are losing market share to R/Python et al. There are tons of free online courses and other resources for both R and Python - just roll your sleeves up and learn them.

Are there extensive, useful, compact (not getting 10% of functions from this library, and 20% from another, etc.) statistical libraries for these languages? I would expect that for R, but am not familiar with what Python has to offer.

Yes you will have to rely on a variety of python libraries to get things done. Part of the skill set is knowing these libraries as well. The good thing is accessing these libraries are made easy using Anaconda.

Quote
Power BI and Tableau are data visualization software. What are you using for data viz?

Right now, I am using IDL for all data analysis. Despite self-promotion, I find IDL to have unfriendly visualization tools. From documentation/help I see that they have made changes to that set over the past 5 years or so, but frankly it is not important enough for my work now to bother exploring that - especially since I won't have the license if I leave the lab, and few other places use it. Plots for papers/grants are made either in IDL or, if simple, Excel. My skill is not in graphic design, but in picking what to present and how, what to correlate with what, what proves/disproves a point, what is relevant to problem at hand, etc.

I checked out Tableau, and I checked out Power BI. While Tableau is on a monthly racket payment plan like so much software, Power BI Desktop is free. Is this good, or too good to be true? If it's a good software package, I would start there, as it would help my research too.

Quote
Kaggle competitions

Looks like fun! It seems like it could be used to learn techniques, I guess that is the point, yes?

Yes, start with something easy like the titanic dataset.

Poundwise

  • Pencil Stache
  • ****
  • Posts: 872
Re: part-time statistics or data analysis gigs - need career advice
« Reply #20 on: August 09, 2017, 02:39:05 PM »
You could try browsing/subscribing to the STAT-L listserv, which occasionally posts stats jobs and courses info. 
http://lists.mcgill.ca/scripts/wa.exe?SUBED1=STAT-L&A=1

As a STEM student, I worked part-time running MANOVA for a consultant who did population forecasts for schools, mall expansions, developers, etc. Small peanuts and probably not at the level you are aiming for, but still a way for a person with stats skills to make a living.

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #21 on: August 11, 2017, 05:48:20 PM »
Quote
If you have a STEM PhD you absolutely don't need a masters, in fact people might question it.

Can you elaborate on this? I would like to get a very good understanding of this before I can nix the MAS idea without later self-doubt and reexamination.

I meant to say during your interview you are going to be asked why you decided to do a masters after your PhD and how convincing you are in your story telling will determine how much they buy your argument. A PhD is considered final and you acquire all the skills you would need to do quantitative research so going back for a masters unless you have a specific goal/story is moot. A lot is going to hinge on that one answer, its tone, its content etc. too much for me personally. I would rather the interview be more focussed on gauging my skills and more positive tone. I interview for my company and in the end the "gut" feeling they walk away with is what matters - along with technical skills ofcourse. So if you are in general a really good sauve communicator I think you'll be ok, if you tend to be even a little bit nervous then the battle will be lost.

...

Yes the STEM PhD should've already taught you how to do quantitative research. Masters starts with knowledge of basic probability theory and algebra, nothing much else is really a prerequisite.

Understood.

milliemchi

  • Bristles
  • ***
  • Posts: 297
Re: part-time statistics or data analysis gigs - need career advice
« Reply #22 on: August 11, 2017, 06:30:56 PM »
There is a lot of talk about some high end statisticery here. I would be grateful if you could just keep some of the yay-whoos I work with out of trouble  with the stuff excel does for them and why a couple data points with a "good enough" R-squared value may not justify millions of dollars in spending on a decision.

As a STEM student, I worked part-time running MANOVA for a consultant who did population forecasts for schools, mall expansions, developers, etc. Small peanuts and probably not at the level you are aiming for, but still a way for a person with stats skills to make a living.

OK, talk to me about yay-whoos and small peanuts, let's explore that.

Poundwise

  • Pencil Stache
  • ****
  • Posts: 872
Re: part-time statistics or data analysis gigs - need career advice
« Reply #23 on: August 15, 2017, 03:52:25 PM »
If you are looking for small peanuts type work, check your local papers for the names of urban planning consultants doing feasibility studies done on school expansions/renovations, etc.  These may be local companies employing only a few people; see if they are looking for help. Warning, consultants live a boom-and-bust life... sometimes there is no work for weeks, then suddenly you are working 80 hours a week for something that needs to be ready yesterday.  But once you have your foot in the door, maybe you can start getting your own clients and making your own hours.

trollwithamustache

  • Bristles
  • ***
  • Posts: 315
Re: part-time statistics or data analysis gigs - need career advice
« Reply #24 on: August 21, 2017, 09:19:08 AM »
6 sigma and now lean 6 sigma has unleashed a horde of the unwashed using excel and sometimes spiffier software to calculate lots of basic statistics and statistical process control stuff. I'm not sure the best way to get in on it, typically you have to go to a couple weeks of classes to get your "green belt" or "black belt". Its best to choke back the vomit during the classes.   

There is some useful stuff but it can turn into a bit of a mess. Typically when 6 sigma goes company wide, everyone with potential gets trained... so there are hourly operator types, one of the smarter secretaries, ect, good people who didn't go to college in the class next to random professionals and engineers.  The pace of the class can't possibly work for all those groups and so it spills out to some learned nothing and the engineers slept through the last hour where they actually needed to refresh on confidence intervals. 

How you sell your services now that these people are "trained" and supposedly productively doing this stuff? that's tricky and I am not sure how, but they need someone to help them, and there are a lot of them out there.

Data quality and understanding if they have any data quality at all will always be an issue with these yay-whoos.