The Artists of Data Science

The ONLY self-development podcast for Data Scientists

Chat Transcript from Data Science Happy Hours 12, Dec 4 2020

16:31:14 From Eric Sims : Ray, your hoodie is awesome!
16:31:22 From Ray Givler : Thanks, man
16:31:30 From Shantanil Bagchi : Congratulations to your sister Harpreet
16:31:38 From Naresh Reddy : Hi Harpreet!
Does it read ''CloaseBy Card"?
16:31:45 From Naresh Reddy : Close*
16:31:46 From Harpreet Sahota : CloseBuy
16:33:01 From Nicole Bills : https://www.closebuyme.com/
16:33:20 From Harpreet Sahota : Thanks Nicole!
16:34:35 From Naresh Reddy : Thanks Harpreet and Nicole!
16:34:59 From Carlos Mercado : Harpreet this shit is so poppin
16:35:48 From Haseeb Mohammed : fireemoji.png
16:36:45 From Haseeb Mohammed : Dialectical behavior therapy (DBT)
16:36:51 From Joe Reis : Data Build Tool
16:37:00 From Haseeb Mohammed : (kidding)
16:37:01 From Haseeb Mohammed : https://www.getdbt.com/
16:37:10 From Joe Reis : lolz
16:37:59 From Dave Langer : AnalyticsOps!!!
16:38:05 From Carlos Mercado : Oh my god. Analytics Ops.
16:38:09 From Eric Sims : @monica - Full stack analytics engineer
16:38:12 From Carlos Mercado : So DevOps, DevSecOps, AIOps, MLOps.
16:38:19 From Haseeb Mohammed : https://blog.getdbt.com/what-is-an-analytics-engineer
16:38:48 From Carlos Mercado : @ Eric lmaooooo dude; I can't. Full stack Analytics Engineer? This is not a real job title.

"Can do SQL and also make nice visuals and also is aware of AWS".
16:38:55 From Mikiko Bazeley : It’s a well -loved tool by Zenefits
16:39:26 From Mikiko Bazeley : My friend Sean is director of analytics there — and they implemented use a couple years ago
16:39:51 From Mikiko Bazeley : Another good post on “analytics engineer”: https://locallyoptimistic.com/post/analytics-engineer/
16:40:07 From Monica Royal : Haha! Yes @Eric... the mystical unicorn
16:40:10 From Timothy Gordon : Great follow-up question Dave. Definitely depends on the environment
16:41:07 From Haseeb Mohammed : the monsters in your head
16:41:08 From Haseeb Mohammed : !!
16:41:51 From Dave Langer : Ben!!!!!
16:41:58 From Ben Taylor : Dave!!
16:43:32 From Ben Taylor : https://techcrunch.com/2016/04/04/saleforce-acquires-metamind/
16:43:46 From Ben Taylor : Price wasn’t public, but someone figured out it was $34M? If I remember right
16:43:56 From Joe Reis : That’s it?
16:43:59 From Ben Taylor : I don’t think they had much revenue
16:44:10 From Joe Reis : Acqui-hire
16:44:23 From Ben Taylor : It was really to get Richard: https://www.linkedin.com/in/richard-socher-19363b18/
16:45:38 From Nicholas Lowthorpe : Hope that's tea in that mug Dave
16:45:46 From Ben Taylor : Question: Why did DataRobot buy Zeff?
16:46:00 From Dave Langer : I'm not telling what's in the mug. ;-P
16:46:01 From Carlos Mercado : Ben shouldn't you be answering that one XD
16:46:32 From Haseeb Mohammed : lol
16:46:32 From Joe Reis : DR bought Zeff cuz of you and Gonzo?
16:48:21 From Nicole Bills : Thank you - that was an awesome debrief
16:48:47 From Shantanil Bagchi : Mikiko..thank you
16:49:11 From Mikiko Bazeley : Totally!
16:50:34 From Carlos Mercado : (1) A GitHub with actual READMEs and a commit history that shows they used GitHub while making the project instead of dumping it on there at the end.
16:51:39 From Mikiko Bazeley : YES!!!!!
16:51:42 From Mikiko Bazeley : LaTex
16:53:01 From Mikiko Bazeley : If it’s on the resume, it’s fair game also
16:53:54 From Matt Housley : I have a love/hate relationship with LaTex
16:54:10 From Mikiko Bazeley : Use LaTex to generate the initial template and then edit is word
16:54:44 From Carlos Mercado : Does LaTeX pass ATS - I assume so, but looking at the backend of LaTeX outputs sometimes, I wonder if it can get busted depending on what program opens it
16:54:45 From Ben Taylor : We’re hiring data science right now… just saying.
16:54:53 From Haseeb Mohammed : ;)
16:55:00 From Ben Taylor : https://www.datarobot.com/careers/
16:55:01 From Carlos Mercado : Ben say it on the audio - you'll get drowned in applications from listeners.
16:55:07 From Sasha Prokhorova : I’ve been advised recently to include some soft skills on the resume, such as leadership, communication etc etc. But for me personally it looks unnecessary and even redundant. What do you guys think? Is it worth it to allocate the valuable real estate to this?
16:55:20 From Mikiko Bazeley : Same Dave
16:55:43 From Carlos Mercado : My opinion on soft skills is you tell business stories that show the skill; don't say "leadership" tell a story that makes it obvious you led.
16:56:08 From Ben Taylor : I have a really funny hedge fund resume story...
16:56:08 From Eric Sims : I wrote a data science slam poem to show I had some understanding of stats and data. It made me stand out, and it worked.
16:56:12 From Carlos Mercado : Like when LinkedIn came out with that article "oral communication is in demand" I was like - nobody talks like that.
16:57:13 From Ben Taylor : I’ve only hired PHDs and a college drop out… I’m missing the middle...
16:57:19 From Ben Taylor : Time for me to hire a BS/MS… :)
16:57:30 From Haseeb Mohammed : working on my masters, dec 2021!
16:57:58 From Eric Sims : Graduating in May and scrolling through the Careers page now :)
16:58:13 From Carlos Mercado : No Eric, you're earmarked for Guidehouse stop
16:58:30 From Eric Sims : Haha, perfect
16:58:33 From DavidTello : I got hired at the Federal Home Loan Bank two weeks before the pandemic and I think what stood out in my resume was a project that I did on interest rates and my prediction on having a recession in 2020. I did the project back on Oct of 2019.
16:59:32 From Carlos Mercado : I'm the 2006 Time Person of the Year.
16:59:36 From Haseeb Mohammed : this
16:59:47 From Eric Sims : @Carlos - Hired.
16:59:48 From Carlos Mercado : https://upload.wikimedia.org/wikinews/en/5/52/TIME_person_of_2006.jpg
17:00:01 From Ben Taylor : Hire slow, fire fast
17:00:10 From Joe Reis : ^ this
17:00:17 From Ben Taylor : If only I could fire people faster… (life goal)
17:00:28 From Joe Reis : Automate jerbs
17:00:33 From Christian Capdeville : show vs tell
17:00:39 From Haseeb Mohammed : https://youtu.be/p1uS5nfIhOg?t=3701

Six things I learned as a Software Engineer trying out Machine Learning - Haseeb Mohammed
17:00:39 From Matt Housley : Async firing?
17:00:48 From Haseeb Mohammed : my chipy talk from earlier this year
17:00:53 From Christian Capdeville : I have a bar graph on my resume. If you can do data viz - show it!
17:01:10 From Ray Givler : ha
17:01:21 From Ben Taylor : Thanks for doing this @Harpreet! Gotta run to a meeting!
17:01:28 From Joe Reis : Later Ben
17:01:30 From Harpreet Sahota : Thanks for coming ben!!!
17:01:32 From Haseeb Mohammed : thanks Ben!
17:02:53 From Nicole Bills : https://towardsdatascience.com/data-science-new-normal-ca34bcbad8f0 is about Carl Gold’s excellent tips for churn analytics
17:03:35 From Carlos Mercado : (I have a small contention with bringing up outliers, it's technically correct, I just worry beginners over-engineer outliers to their detriment) but its a good article
17:03:59 From Nicholas Lowthorpe : The bit about including projects in your resume is that it gives you an opportunity to focus on a result and not a process
17:04:21 From Christian Capdeville : Very similar to the classic elevator pitch format: "I help X with Y by doing Z"
17:04:39 From Nicole Bills : re: outliers - I think beginners over-engineer a lot of data science preprocessing efforts
17:04:55 From Haseeb Mohammed : the oldest stuff on my resume is summed up in a single sentence. my recent stuff is more spelled out for each project
17:05:35 From Carlos Mercado : I love to use catch phrases on my resume. I have "Technical Sales lead for the AI group - win big work more often". I was product manager overseeing a document summarization tool and said "hours of reading, in minutes".

Throw in copywrite, it's a whole skill and it shows communication skill.

17:05:48 From DavidTello : I would limited showing work experience to no more than 10 years and eliminate anything that is not that relevant to the position you are applying for
17:06:06 From Nicholas Lowthorpe : +1 for copywriting Carlos
17:06:33 From Haseeb Mohammed : anomaly detection
17:06:37 From Haseeb Mohammed : cascading learning
17:06:40 From Haseeb Mohammed : SMOTE
17:06:45 From Haseeb Mohammed : undersampling
17:06:47 From Haseeb Mohammed : oversampling
17:08:07 From Greg Coquillo To Harpreet Sahota(privately) : I do have a question
17:08:07 From Haseeb Mohammed : "Learning from Imbalanced Data Sets"
17:08:16 From Carlos Mercado : (1) Verify it's actually a problem. 20%, probably not a problem to be honest. 3% - that's where things get tough. Fraud detection especially.
(2) What makes sense in the business context (false positives vs false negatives have costs and your options will mess with this balance)
(3) then do the mathy stuff and test it out on a training sample.

(3b) split test/train/validate prior to re-sampling. This is something people actually forget.... Posted an article about it.
17:09:27 From Carlos Mercado : There are also situations where resampling works on training data and is useless in production. You don't get to resample in production. So again, (1) verify it's actually a problem.
17:09:31 From Timothy Gordon : Appreciate the responses everyone!
17:10:08 From Harpreet Sahota To Greg Coquillo(privately) : Ok for sure!
17:10:15 From Haseeb Mohammed : "There are also situations where resampling works on training data and is useless in production. You don't get to resample in production."
17:10:18 From Haseeb Mohammed : 100%
17:10:28 From Haseeb Mohammed : in training you can be really happy with your oversampled/undersampled dataset
17:10:49 From Timothy Gordon : Can you share the article on 3b Carlos?
17:11:12 From Sasha Prokhorova : Lol skill bubbles :))
17:11:40 From Haseeb Mohammed : someone said it above, everything on the resume is fair game
17:12:04 From Haseeb Mohammed : during our software engineering interviews, i pick out the obscure stuff on the resume and ask them why they added it to their resume, what they've done with it.
17:12:16 From Carlos Mercado : Amen ^
17:12:18 From Haseeb Mohammed : i loathe the skills section, when it doesn't include the project you used those skills on
17:12:21 From Shantanil Bagchi : Thanks Haseeb
17:12:26 From Mikiko Bazeley : David’s comment on real estate pleases me
17:12:32 From Sasha Prokhorova : What if a person is trying to show the bubbles for the comparison between the skills? For instance, they are more comfortable with SQL than Python etc.
17:12:44 From Carlos Mercado : nah
17:12:57 From Carlos Mercado : you're either ready to create value with the skill or not. It's binary.
17:13:03 From Timothy Gordon : Great comment Haseeb definitely want to highlight your abilities with actual proof
17:13:14 From Haseeb Mohammed : you're coming in for a junior position, i'm going to assume you can't do anything anyway -- software related
17:13:40 From Nicholas Lowthorpe : I had a 45 minute conversation with a data and analytics recruiter the other day about the state of recruitment. We had a half-joking conversation about the use of numbers to skills, where we concluded that the only ranking out of 5 you can realistically give yourself is a 4. If you list yourself 3 out of 5 you're saying you're average, and you'd never write on your CV "average at python". 5 is essentially saying 'mastery' because there's no room to improve. Anything less than 3 is a skill you wouldn't list on your CV. So you'd list everything at 4!
17:14:00 From Nicholas Lowthorpe : (hence don't use numbers)
17:14:05 From Carlos Mercado : People actually say "average" "beginner" I hate it, why say "I suck" on a sales document!
17:14:14 From Carlos Mercado : Resumes are a sales document.
17:14:16 From Haseeb Mohammed : https://i.imgur.com/r0SEEoh.jpeg
17:14:21 From Shantanil Bagchi : Nicely said NIcholas
17:14:23 From Mikiko Bazeley : Know what you’re selling with your resume
17:14:26 From Mikiko Bazeley : period
17:14:30 From Carlos Mercado : Amen ^
17:14:33 From Haseeb Mohammed : 100%
17:14:35 From Christian Capdeville : Solid points Nicholas and carlos
17:14:41 From Mikiko Bazeley : Are you selling features or a solution?
17:15:52 From Carlos Mercado : RE: domain expertise
Read case studies - like 3-8 page things. You want to have a bunch of stories.

Examples:
IBM/Walmart Blockchain
DataRobot/Covid-19
RStudio - Blog
17:16:20 From Joe Reis : I heard some apocryphal tale that Guido Van Possum (the creator of Python) rated his Python skills 6/10. So there’s that.
17:16:26 From Joe Reis : Rossum
17:16:27 From Carlos Mercado : haha
17:16:30 From Haseeb Mohammed : lol
17:16:44 From Monica Royal : That is great advice @Nicole!
17:17:11 From Dave Langer : These days you can also articles, books, and online courses that are dedicated to analytics to particular business domains.
17:17:20 From Nicholas Lowthorpe : Nicole Bills - that is absolutely brilliant. I literally just googled and found my local authority has a ton of public data for anonymised people movement through mobile cell towers and air quality
17:17:25 From Nicholas Lowthorpe : Awesome advice
17:17:52 From Dave Langer : For example:
17:17:53 From Dave Langer : https://www.amazon.com/Marketing-Research-Analytics-Use/dp/3030143155/ref=sr_1_3?dchild=1&keywords=R+marketing+analytics&qid=1607123848&sr=8-3
17:18:37 From Nicole Bills : lol I like Van Possum
17:19:32 From Joe Reis : Autocorrect yields a cooler Guido
17:19:40 From Haseeb Mohammed : i leave my only shiny nugget of advice for resumes, no full sentences
17:20:07 From Nicole Bills : Amazon does like the STAR method
17:20:53 From Nicole Bills : Their 14 Leadership Principles are kind of cult-y, but effective
17:21:02 From Carlos Mercado : If you haven't watched the AlphaGo movie:
https://www.youtube.com/watch?v=WXuK6gekU1Y
Alpha stuff is crazy
17:21:39 From Shantanil Bagchi : Yah..it really mindboggling...the top player was so shocked initally
17:21:42 From DavidTello : I think it matters if the sample space was truly random @Greg
17:21:48 From Carlos Mercado : I cried during the movie for real
17:24:15 From Brandon Quach : This is great! I learned so much from all the experts on this call. I’ve got to get back to work (Pacific Time. California).
17:24:16 From Nick Urban To Harpreet Sahota(privately) : Congrats, biggest OH I’ve seen!
17:24:29 From Brandon Quach : Bye everyone!
17:24:35 From Harpreet Sahota To Nick Urban(privately) : Insane
17:24:36 From Shantanil Bagchi : Thanks Brandon
17:24:36 From Eric Sims : Bye, Brandon!
17:24:36 From Austin Loveless : Bye Brandon! Have a good one
17:26:09 From Haseeb Mohammed : ive got to jet as well, take care folks!
17:26:12 From Carlos Mercado : @ Mark
https://12factor.net/
17:26:17 From Haseeb Mohammed : thanks Harpreet!
17:26:24 From Carlos Mercado : Fundamentals of products is fundamentals of software engineering.
17:26:38 From Austin Loveless : Have a good one Haseeb!
17:26:54 From Carlos Mercado : Specific to your question Mark: https://12factor.net/codebase
17:27:23 From Carlos Mercado : To maybe speak for Dave - I think the answer is "Fast" first; Scale when it hurts.
17:27:41 From Ray Givler : Gotta go. Thanks everyone!
17:27:50 From Eric Sims : See you, Ray!
17:29:27 From Joe Reis : Later Ray
17:30:06 From Carlos Mercado : There's a difference between Open Source & Free and Open Source - learning that the hard way
17:30:13 From Eric Sims : Dave, you ask great clarifying questions! It's really helpful.
17:30:33 From Carlos Mercado : https://opensource.com/article/17/11/open-source-or-free-software
https://www.zdnet.com/article/foss-vs-open-source-as-an-american-debate/#:~:text=Is%20IBM%3F-,Open%20source%2C%20in%20contrast%20to%20FOSS%2C%20accepts%20the%20idea%20that,other%20words%2C%20is%20Irving%20Berlin.
17:30:50 From Timothy Gordon : ^ agree great questions Dave! Learning a lot from everyone here
17:31:49 From Carlos Mercado To Harpreet Sahota(privately) : Harpreet, this has been a sick one. I think a small tweak would be including a separate LinkedIn post or email address in the videos for people to ask questions asynchronously to bring up here.
17:32:14 From Harpreet Sahota To Carlos Mercado(privately) : Yeah - good point
17:33:14 From Nicole Bills : Brent

17:33:23 From Carlos Mercado : Thank you! Brent*
17:33:24 From Jacqueline Lefèvre López : I've got to go, thank you for all the advice :) this was great!
17:33:28 From Carlos Mercado : I just read it too LOL
17:33:39 From Dave Langer : Joe - BTW, I have supported a DB2 system on Z-Series at one point back in the 1800s
17:33:55 From Austin Loveless : Have a good one Jacqueline!
17:34:27 From DavidTello : COBOL :)
17:34:34 From Dave Langer : Hold on, COBOL is not dead!
17:34:41 From DavidTello : Joking
17:34:46 From Dave Langer : :-)
17:34:48 From Joe Reis : dave'
17:34:52 From Carlos Mercado : 01000011 01001111 01000010 01001111 01001100 00100000 01110011 01110101 01100011 01101011 01110011
17:34:57 From Joe Reis : Has been writing Cobol since 1868
17:35:07 From Sasha Prokhorova : I had a podcast episode on COBOL not so long ago - it appears to be very much alive
17:35:25 From Dave Langer : Those were the good old days. 6-shooters, snake oil, and COBOL!
17:35:35 From Mark Freeman : That was extremely helpful!!! Thanks everyone!
17:35:37 From DavidTello : please share the URL, I would like to listen to it @Sasha
17:35:56 From Venkataramana Maram : samr here with ...
17:36:12 From Venkataramana Maram : same here with
17:37:33 From DavidTello : Don’t take that for granted @Sasha. Most programmers/DS are seem as introverts
17:37:55 From Joe Reis : You can show your communication skills by giving talks (and getting them recorded)
17:38:13 From Mikiko Bazeley : I was thinking with covid
17:38:17 From DavidTello : Try Toastmasters, it helped me a lot
17:38:24 From Carlos Mercado : I gotta bounce, but thank you all, these are so great!!
17:38:31 From Joe Reis : Later carlos
17:38:31 From Eric Sims : Later, Carlos!
17:38:38 From Nicole Bills : Bye Carlos!!
17:38:39 From Timothy Gordon : seconding David with Toastmasters
17:38:44 From Naresh Reddy : I got a qucik question, Harpreet!
17:38:51 From Austin Loveless : By Carlos
17:38:58 From Austin Loveless : Bye*
17:39:34 From Nicholas Lowthorpe : Anyone who's interested in toastmasters and might want to take it to the next level - there is a global public speaking competition in STEM called FAMELAB - runs every year
17:39:34 From Eric Sims : This is a chat message Easter Egg for the NLP challenge. Message me on LinkedIn if it's 2021 and you found this...
17:39:43 From Nicholas Lowthorpe : it's a lot of fun
17:39:52 From Mikiko Bazeley : Hey Naresh could you speak up
17:40:01 From Nicole Bills : Data viz based on transcript NLP
17:40:21 From Mark Freeman : I have to get back to work. Always appreciate the knowledge shared here!
17:40:29 From Joe Reis : Later Mark
17:40:38 From Nicole Bills : Going to hop offline as well - thanks all!
17:40:40 From Austin Loveless : Have a good one Mark!
17:40:49 From Joe Reis : Later Nicole
17:40:53 From Austin Loveless : Later Nicole! Appreciate your inisights!
17:40:58 From Austin Loveless : insight* wow haha
17:41:37 From DavidTello : Thank you @Nicholas, I never heard of FAMELAB before, I will check it out
17:42:06 From Timothy Gordon : Haven't heard of FAMELAB either thanks!
17:42:09 From Nicholas Lowthorpe : There are heats in most places - the format is 3 minutes, no notes, talk about a topic to an audience and a panel, Q&A after
17:42:25 From Austin Loveless : Same! I'll check it out Nicholas :)
17:43:02 From Nicholas Lowthorpe : I took part twice, I progressed to Northern UK and won my first time, then the year after I got too confident and went down in a ball of flames in round 1 :)
17:45:24 From Joe Reis : If you’re in college, join something like the debate team. I found debate super useful for developing my communication skills
17:46:27 From Joe Reis : Also, Warren Buffett says one of his best investments was a public speaking course
17:46:28 From Joe Reis : https://www.cnbc.com/2017/10/04/warren-buffett-says-this-one-investment-supersedes-all-others.html
17:49:24 From DavidTello : Wolf of Wall Street
17:52:09 From Sasha Prokhorova : I could use some public speaking classes myself!
17:53:10 From Mikiko Bazeley : I swear I got that exact line from my early business partners as feedback on my presentations
17:53:25 From Dave Langer : More Meetup talks, less deep learning, people!
17:53:28 From Dave Langer : ;-P
17:53:30 From Matthew Blasa : Yup. My boss told me once not to puke data.
17:53:35 From Joe Reis : yep
17:54:09 From Venkataramana Maram : pls write in chat ...Harpreet for books
17:54:12 From Joe Reis : Also, read “never split the difference”
17:54:31 From Nicholas Lowthorpe : robert cialdini - pre-suasion
17:54:33 From Nicholas Lowthorpe : was one of them
17:54:34 From Matthew Blasa : Meditations has a graphic novel?
17:54:35 From Nicholas Lowthorpe : awesome book
17:54:36 From Joe Reis : And influence
17:54:48 From Mikiko Bazeley : Crucial Conversations
17:54:49 From Austin Loveless : The art of selling anything was another I believe
17:54:55 From Austin Loveless : To sell as human I think the third?
17:54:55 From Joe Reis : I ran into Robert last year. A new version of Influence is coming out soon
17:54:58 From Venkataramana Maram : great
17:55:03 From Venkataramana Maram : thanks

Article Comments