The Artists of Data Science

The ONLY self-development podcast for Data Scientists

Chat Transcript from Data Science Happy Hours 22, March 5, 2021

16:31:11 From Naresh Reddy : Hello everyone
16:31:27 From Nicholas Lowthorpe : hey guys
16:31:29 From Russell Willis : Evening All
16:31:31 From Eric Sims : Yo!
16:31:32 From Nicholas Lowthorpe : and gals
16:31:44 From Ashit Debdas : Hello everyone
16:31:56 From Nicholas Lowthorpe : we've got Albert Bellamy LIVE from the (MAU) mobile analytics unit
16:32:18 From Eric Sims : @Albert with the #48HoursofData coming up!!!
16:32:43 From Eric Sims : @Rodney - Great to see you!
16:33:54 From Russell Willis : Gelotologist, or JelloShotologist..., I think I'd be cool with learning more about either ;)
16:34:36 From Carlos Mercado : Normal, Poisson, gamma, beta, chi2, Weibull (exponential). But a lot of distributions are different scaled versions of other distributions
16:34:37 From Rodney Beard : @Eric hi!
16:34:49 From Eric Sims : Thanks, @Carlos!
16:35:06 From Mikiko Bazeley : I feel like it’s also dependent on domain — when I took supply chain, poisson was king
16:35:11 From Carlos Mercado : agreed^
16:35:20 From Carlos Mercado : A/B tests use Weibull a lot
16:35:25 From Akshay Mandke : https://towardsdatascience.com/a-field-guide-to-the-most-popular-parameters-d734596c3f26
16:35:37 From Albert Bellamy : I was gonna say, no love for the Weibull???
16:35:40 From Rodney Beard : Kotz and johnson classic ref for distributions
16:35:44 From Akshay Mandke : Cassie Kozyrkov recently shared a post on distributions
16:35:51 From Carlos Mercado : Saw but didn't read that ^
16:36:01 From Thomas Ives : Sorry to cut you off Carlos!
16:36:13 From Russell Willis : @harpreet What about more basic Statistical distributions single point / 3 point , PERT, Mean, etc?...
16:36:24 From Thomas Ives : Albert! Thanks for serving brother!
16:36:29 From Rodney Beard : Triangular is terrible
16:37:33 From Carlos Mercado : Bernoulli!
16:37:57 From Carlos Mercado : Is the IBM Data Science Professional cert legit? I'm hearing some in industry disproportionately like it.
16:37:57 From Mikiko Bazeley : I think it’s good to intervene if cost is an issue as well
16:38:16 From Thomas Ives : Yes Mikiko - AGREE!
16:38:36 From Mikiko Bazeley : If someone is asking about a $500-600$ cost class, I like to point to something more available
16:38:47 From Carlos Mercado : Oh fair point^
16:39:14 From Russell Willis : Don't try to understand ALL distributions in one bite... there are tangible relationships between some, so they can lend themselves to an organic learning journey...
16:39:16 From Carlos Mercado : I liked the Coursera Data Science Specialization (~$500). It's how I got started.
16:39:17 From Thomas Ives : Maybe take a little time to tell them there's no perfect course and that they need to focus on concepts and to keep a living ever growing learning plan.
16:39:19 From Harpreet Sahota : If you have a question - let me know! I will add you to the queue
16:39:31 From Eric Sims : Oof, that's expensive!
16:39:36 From Thomas Ives : Love your point Russell!
16:40:16 From Carlos Mercado : Eric - universities are like $60k? LOL
16:40:39 From Thomas Ives : If it's good, $500 is cheap compared to Uni's
16:40:47 From Eric Sims : Haha, yeah, but mine has a much better job placement rate than a Udemy course :)
16:41:12 From Carlos Mercado : NCSU has the actual best job placement numbers and report of any analytics program I've ever seen.
16:41:19 From Thomas Ives : Good point Eric.
16:41:22 From Greg Coquillo : I have a question
16:42:10 From Thomas Ives : Vikram!
16:42:44 From Vikram Krishna : Thom !!
16:43:07 From Ben Taylor : Levy distributions are my favorite..
16:43:11 From Kristen Davis : It’s tough though, as someone entering the field the entry level jobs out there are asking for masters or phd - seems a disconnect from the community which really embraces the power of self learning over the degree and these hiring managers emphasizing the degree not the passion / self learning
16:43:17 From Carlos Mercado : Thomas - you ever do that computer vision thing with the student who called in a few months ago?
16:43:19 From Eric Sims : Ooh, the dictionary is a really good comparison! I like that
16:43:33 From Mikiko Bazeley : I did boot camps & a program — so definitely not knocking paying high value programs & workshops (Reforge was $3K for a workshop) but I always like gauge commitment before ending them down the $$$ path
16:43:44 From Eric Sims : ^
16:44:00 From Mikiko Bazeley : But going to bootcamp was the best thing for me and no regret
16:44:14 From Carlos Mercado : I'm dying bro. This answer is amazing.
16:44:17 From Austin Loveless : Glad to be back :) have had the last few weeks where I had something scheduled. Just eating atm so will hop on camera in a bit!
16:44:17 From Eric Sims : I took free and inexpensive courses before committing to a degree. My wife actually talked me into going back to school
16:44:27 From Thomas Ives : Carlos, I tried to hang in there with them on LinkedIn messaging, and we talked for a while, but then I stopped hearing from them.
16:44:40 From Carlos Mercado : dang, at least you tried!
16:44:57 From Saurabh Dixit : Which bootcamps are good.. simple to grasp and building good foundation
16:45:00 From Christian Capdeville : Vin - this hits so close to home it hurts. But even if someone would've told me, I probably would have ignore it
16:45:04 From Tor Rorvik : Questioning the PPM (Power Point Money)
16:45:17 From Ben Taylor : I wish I had been told that job security is a lie, you need to fight for market security
16:45:18 From Vikram Krishna : @Saurabh DPhi are good
16:45:24 From Eric Sims : Reality 1010 sounds like a tough class...
16:45:37 From Russell Willis : @Eric "the distribution of my course has fewer outliers, when it comes to positive job placements!..."
16:45:46 From Mikiko Bazeley : Wish they taught that in high School lol
16:45:55 From Vin Vashishta : Thanks for the book reco.
16:46:00 From Eric Sims : Amen, @Mikiko!
16:46:08 From Thomas Ives : Good point Ben
16:46:20 From Saurabh Dixit : Thanks @Vikram, @Eric
16:46:22 From Russell Willis : +1 @Ben
16:47:21 From Eric Sims : Ordering a BS detector on Amazon right now...
16:47:35 From Carlos Mercado : David's been a good counter to all my blockchain bs.
https://www.google.com/search?q=the+art+of+skepticism+in+a+data-driven+world
16:47:39 From Akshay Mandke : +1 prime order
16:47:46 From Russell Willis : +1 @David BS and unnecessary hyperbole, both!
16:47:53 From Eric Sims : lol, @carlos, I love it!
16:48:04 From Ben Taylor : Thomas’s book video was legendary...
16:48:09 From Carlos Mercado : I linked the google search so that people didn't necessarily buy from Amazon lol
16:48:19 From David Knickerbocker : Sorry, yes, "Calling Bullshit"
16:48:42 From Mikiko Bazeley : Love that web comic
16:48:56 From Mikiko Bazeley : http://phdcomics.com/
16:49:16 From Nicholas Lowthorpe : I love that advice Thom.
16:49:30 From Russell Willis : +1 Thom!
16:49:41 From Saurabh Dixit : Great message Thomas
16:50:16 From Ashen Rana : Learning Data Science vs working in data is eye opening. Need to get your feet wet in data world somehow and learning becomes less intimidating
16:50:31 From Ashen Rana : Add value = job security
16:50:38 From Carlos Mercado : My uninformed opinion on PhDs.
http://phdcomics.com/comics/archive.php?comicid=1286
16:50:43 From Mikiko Bazeley : Something about penetration of ML in production: slide 40 here: https://www.slideshare.net/sergeykarayev/lecture-5-ml-projects-full-stack-deep-learning-spring-2021?ref=https://fullstackdeeplearning.com/
16:50:49 From Eric Sims : Hey Ben! Enjoyed your DataTalks episode with Alexey today 👍
16:50:55 From Mikiko Bazeley : Still lots of opt for people to enter the field
16:50:56 From Vikram Krishna : I started my Data Science journey with "Python Data Science Handbook"
16:51:02 From Christian Capdeville : Nice Thom, reminds me of: Resourcefulness vs Resources - this comes up everywhere. If you don't have the resources - good news! You're about to learn how to be very resourceful...
16:51:31 From Austin Loveless : Build that social proof by learning in public: https://www.swyx.io/learn-in-public/
16:51:43 From Carlos Mercado : Austin I still post this link like 3x a day
16:51:50 From Robert Robinson : Most hiring managers don't really understand what they are looking for.
16:51:51 From Mikiko Bazeley : It’s the False Positives companies are trying to defend against
16:51:54 From Mikiko Bazeley : Supposedly
16:52:02 From Mikiko Bazeley : Is why they fall back on MS/PhD
16:52:14 From Austin Loveless : It's a great resource @Carlos. Glad to hear you're still spreading the word
16:52:40 From Russell Willis : @Thom - There has been some good posts on LinkedIn recently about unorthodox and meandering career paths often leading to wider and more resilient skillsets. @BenTaylor has been active in some of these
16:52:54 From Thom (like Thomas without the ass) Ives : Agreed Robert!
16:53:50 From Rodney Beard : Measurement theory
16:53:58 From Rodney Beard : For building intuition
16:53:59 From Carlos Mercado : Echoing Khuyen - I was on 0 people's radar prior to LinkedIn. I think personal branding is just printing value right now. It's still not fully saturated. New platforms coming out regularly and ways to leverage multiple at once.
16:54:03 From Mikiko Bazeley : Packt has a bunch of those “DS/ML for…..” books
16:54:18 From Tor Rorvik : Don't forget that 1st round Application/ resume review is performed by Mr/ Mrs Algorithm and not a human being.... now a days..... hence important to have the right "SEO" wording in there.... 2nd round review is human, that's hen you need projects and a good storyline and then the interview.....
16:54:24 From Thom (like Thomas without the ass) Ives : Greg, That just comes with doing real world projects over time and thinking a lot and doing tons of visualizations throughout the process.
16:54:34 From Carlos Mercado : Packt asks me every week if I'll review a book and then they send me a copy and then I don't read it and then they destroy my inbox for a review
16:54:52 From Mikiko Bazeley : Reforge also has some workshops focused on experimentation & testing for Growth/product
16:54:53 From Thom (like Thomas without the ass) Ives : Good point Tor!
16:55:02 From Harpreet Sahota : If you have a question, let me know! I will add you to the list
16:55:14 From Carlos Mercado : @ Tor - Resumes are sales documents. Agreed on SEO.
16:55:17 From Eric Sims : +1 to Carlos' statement (about making value, not about Packt destroying his inbox...)
16:55:20 From Vikram Krishna : @Carlos I am in a similar situation right now!
16:55:37 From Jaya To Harpreet Sahota(privately) : I have a question from the Sunday mentoring
16:55:50 From Mikiko Bazeley : I feel attacked by the “read then do nothing” cmment
16:55:53 From Mikiko Bazeley : XD
16:55:58 From Tor Rorvik : So instead of saying: I have 3 years of experience, say: I feel like I have done this for 25 years......
16:56:05 From Eric Sims : lol
16:56:17 From Russell Willis : @Greg Do you mean to use past output accuracies to determine which distributions to apply first, rather than trial and error?
16:56:53 From Carlos Mercado : YouTube is the actual GOAT resource for applied learning. You just copy them, then randomly mix it up and experiment with the code, then watch other videos that explain stuff. I could be on YouTube 10 hours/day without even thinking about it.
16:57:03 From Carlos Mercado : But yeah it's not structured to Vin's point.
16:57:27 From Thom (like Thomas without the ass) Ives : I love that Carlos!
16:57:40 From Nicholas Lowthorpe : The idea of learning from YouTube fills me with dread. I hate learning from video, unless there's a transcription I can read.
16:57:48 From Rodney Beard : Measurement theory works really well for building the link between application and distributions. A lot of it is in introductory stats books but is often not mentioned by that name, but that’s where it comes from. Examples nominal scaled, interval scaled, ratio scaled variables, count data etc. these map well to distributions although imperfectly.
16:58:01 From Thom (like Thomas without the ass) Ives : Nicholas 😂
16:58:11 From Nicholas Lowthorpe : Give me a textbook any day.
16:58:18 From Greg Coquillo : @Russell yes, if you're new in the field, how do you tackle business cases you've never seen before? How do you do that with high intuition for building the bridge between a use case and probability distribution?
16:58:26 From Thom (like Thomas without the ass) Ives : YouTube and Popular Mechanics has saved my ass at times!
16:58:37 From Mikiko Bazeley : Daniel Bourke has a great idea about the reverse kaggle
16:58:42 From Carlos Mercado : To Vin's point about the problem for bootcamps (resources) vs colleges (inflexible). The best scenario is (and was prior to like the 1990s) that corporations do the training as close to the daily work as possible.
16:59:08 From Mikiko Bazeley : Reverse Kaggle: https://www.linkedin.com/feed/update/urn:li:activity:6770978116107161600/
16:59:10 From Eric Sims : There is no safe space. I have found I can be proven wrong pretty much everywhere... :)
16:59:30 From Austin Loveless : Thanks Mikiko I was just looking for that
16:59:38 From Vivienne DiFrancesco : I'd be curious to hear from any of the experienced people about your "oh crap" moments. Are there times when you did some project and later realized that you made mistakes after implementation was already done or underway?
17:00:16 From Carlos Mercado : Google's academy for google suite, Amazon learning paths for AWS, etc. are the structure we should aim for, in my opinion. This is actually where community colleges can fill in a lot of gaps. Less overhead, more relationships to community jobs, etc.
17:00:18 From Russell Willis : I think that using "real-world" data, including anomalies and errors is some of the best training around! if training exclusively with precleaned data sets, then the "real-world" transition will be PAINFULL!!
17:00:59 From Carlos Mercado : I like real-world; but I also like to recommend people learn data simulation.
17:01:08 From Austin Loveless : Data simulation?
17:01:11 From Carlos Mercado : Making random data that has patterns to mine is a tough problem.
17:01:38 From Ben Taylor : I’m walking away for 5-10min… I’ll remain on the call. Helping throw my 6 yr old in the car so he can go to brain balance.
17:01:38 From Rodney Beard : Simulation is really useful
17:01:39 From Wilson Man : Sweet, good question to enter on.
17:01:46 From Vin Vashishta : My earliest oh crap moment was when I realized how hard sourcing data is. Not so much real world data, etc. but what happens when the business does not have the data and does not have access to the data?
17:02:09 From Christian Capdeville : Random crossover question (applies to both data projects and product creation) - do you have any favorite questions to ask users to find out what they actually want/need (rather than what they are asking you for)?
17:02:10 From Russell Willis : one counter to this is the potential to develop data trusts, in which to coalesce and aggregate multiple sources of data, for increased learning pools... However, this needs buy in form multiple parties, to be prepared to give up their data, without fear of losing IP, or other market advantage...
17:02:23 From Thom (like Thomas without the ass) Ives : Creating Fake data is a powerful skill too IMO.
17:02:28 From Eric Sims : @Austin - I just learned today that Pandas has a built in dummy data generator!
17:02:34 From Nicholas Lowthorpe : Vivienne - I imaged assault rifles in luggage to create a small dataset resembling x-ray images (think airport baggage scanners). To flesh it out, I then used that to create tons of artificial imagery. My CNN ended up being really good at identifying photoshopped items in x-ray images...
17:02:37 From Carlos Mercado : Something I was telling a junior consultant who joined my team was, "trust, but verify". Never trust data collectors to collect in a way that is amenable to analysis. Get involved as close to data generation as possible. Data simulation helps with this skill.
17:03:09 From Saurabh Dixit : Can you repeat the book title please “Pragmatic Programming… ?
17:03:23 From Thom (like Thomas without the ass) Ives : Agreed Carlos!
17:03:25 From Harpreet Sahota : And pragmatic thinking and learning
17:03:30 From Austin Loveless : @Eric, oh wow I didn't realize that. I only knew about the "get_dummies" for one hot encoding haha
17:03:34 From Saurabh Dixit : Got it.. thanks
17:03:53 From Thom (like Thomas without the ass) Ives : It's a process Jaya!
17:03:55 From Russell Willis : -- data trusts will be entirely reliant upon being able to effectively anonymise/pseudonymise multiple data sets, so as not to allow reverse engineering of data to be identifiable from any single contribuiting entity!!
17:04:04 From Thom (like Thomas without the ass) Ives : Yes Harpreet! MUST 1st step!
17:04:11 From Wilson Man : Didn't know about the dummy data generator either. Thanks Eric.
17:04:14 From Eric Sims : @Austin, this article talks about it in point 2: https://realpython.com/python-pandas-tricks/
I think it still exists and hope it hasn't been deprecated.
17:04:19 From Thom (like Thomas without the ass) Ives : Then STAY engaged
17:04:33 From Austin Loveless : Trust the data initially as far as you can throw it.
17:04:54 From Austin Loveless : Thanks @Eric. I dig the resource
17:05:04 From Russell Willis : +1 @Thom creating artificial data for testing purposes, is invaluable!
17:05:14 From Carlos Mercado : RE: convincing managers to start data projects.

(1) Are you sure data is needed to generate value? You can generate value with easier things like Excel templates, RPA bots, and process improvements and it's almost guaranteed value. Data might not generate value.

(2) Do you know anything about the data they 'hold close'? Are you sure it's valuable? You might be way too early
17:05:27 From Vivienne DiFrancesco : +1 @ Nicholas
17:05:40 From Carlos Mercado : I think data CAN help companies, but they might not be collecting anything useful.
17:06:16 From Austin Loveless : Agreed @Carlos. Sometimes the value is actually trying to augment/enhance a process.
17:06:27 From Harpreet Sahota : If you have a question let me know, I will add you to the queue
17:06:37 From Wilson Man : Kind of on the same train of thought as Carlos. Data and Excel are not mutually exclusive, so I'm not entirely sure that they're not using data, just because they're using Excel.
17:06:41 From Thom (like Thomas without the ass) Ives : Spot on Mikiko!
17:06:43 From Carlos Mercado : Our Robotic Process Automation team has grown faster than our Analytics team for over 2 years.
17:06:45 From Akshay Mandke : add to queue
17:07:00 From Harpreet Sahota To Akshay Mandke(privately) : added
17:07:10 From Tor Rorvik To Harpreet Sahota(privately) : i like to respond
17:07:19 From Harpreet Sahota To Tor Rorvik(privately) : ok
17:07:48 From Russell Willis : RPA is a huge growth area!
17:08:25 From Mikiko Bazeley : Getting buy-in isn’t always a 1-shot thing — sometimes it needs to be built over small increments
17:09:16 From Russell Willis : +1 @Carlos, though create them an Excel template, then wait 10 seconds for them to break it!!
17:09:21 From Harpreet Sahota To Vivienne DiFrancesco(privately) : Just saw your Q I will add you in
17:09:43 From Wilson Man : Right in the feels, Russell. :'(
17:09:59 From Mikiko Bazeley : Gsheets >>> Excel
17:10:03 From Austin Loveless : A huge time saver that recently came from our finance group was automating a weekly KPI report that would take HOURs each week to pull from various data sources.

They finally pulled the data into the Business Intelligence tool and automated the exporting of the spreadsheet to go out to a list every Monday.
17:10:22 From Russell Willis : +1 @Mikiko... Especially if it is a big shift from existing paradigms!
17:10:33 From Thom Ives : Nice Russell
17:10:36 From Mikiko Bazeley : It’s a nice way to get people bought into the cloud
17:11:01 From Eric Sims : If you want an awesome no-code/low-code tool for process automation across different platforms, check out Zapier. It's my fave.
17:11:09 From Mikiko Bazeley : At one company we built a bunch of dashes out of sheets and some plug-ins
17:11:15 From Christian Capdeville : Recurring theme around how to convince people of stuff: Focus on outcomes (for them - not you)
17:11:20 From Carlos Mercado : "A lot of customer complaints" LOL
17:11:31 From Thom Ives : Think of your data science efforts in a company as being a startup, and do customer centric design for your internal customers.
17:11:36 From Eric Sims : @Christian - BAM! Yes.
17:11:50 From Mikiko Bazeley : THE Eric Sims
17:11:52 From Saurabh Dixit To Harpreet Sahota(privately) : Hi, I need to drop off now.. but it was amazing to join and hear those gems. It’s amazing to see so many awesome books that you have on the list / recommendations. Great to catch up. Good day and take care.
17:11:58 From Mikiko Bazeley : mic drop
17:12:00 From Thom Ives : Love that Christian!
17:12:06 From Carlos Mercado : A really basic thing you can do is study their process and identify the "simplest data entry". If there is something that is ALWAYS the same and they are just entering data? Can you turn that into a structured google/access form?
17:12:16 From Thom Ives : Eric, Accept your fame and popularity!
17:12:24 From Christian Capdeville : Eric is the man!
17:12:24 From Thom Ives : U Duh Mahn!
17:12:28 From Eric Sims : Haha :)
17:12:46 From Mikiko Bazeley : All about performance over baseline
17:12:54 From Ben Taylor : I love Mikiko’s toy story background
17:13:03 From Mikiko Bazeley : <3
17:13:05 From Carlos Mercado : Get that used to the idea that "Excel has limits". Because right now they have PROOF that Excel = profit. You need to build proof that Not Excel = MORE profit.
17:13:11 From Carlos Mercado : get them used*
17:13:47 From Vivienne DiFrancesco To Harpreet Sahota(privately) : Cool, thanks!
17:14:09 From Mikiko Bazeley : The Challenger Sale is a great book about bringing people to the precipice of fear
17:14:22 From Mikiko Bazeley : And then selling them an enterprise solution
17:14:27 From Tor Rorvik : INTERNAL AUDIT RISK - FREQUENCY and CONSEQUENCE = RISK MATRIX
17:14:49 From Christian Capdeville : Absolutely Vin - Incentives are everything
17:15:00 From Russell Willis : For Scrappy solutions, one potential ugly truth is that to improve, may entail changing it and monitoring the shift, to keep positive shifts and undo negative shifts... Guerrilla improvement!
17:15:48 From Nicholas Lowthorpe : 2 marketing mantras:

  1. "sell the benefits, not the features" is key for getting your idea across the line. If you've identified a problem, then present your solution as a benefit that is as close to their world as possible. Revenue, EBITDA, etc.

  2. "people will walk to pleasure but run from pain" - benefits are a great sell, but often a stronger sell is to talk about the pain someone will encounter if they don't act.
    17:15:54 From Nicholas Lowthorpe : They're highly applicable
    17:16:10 From Russell Willis : +1 @Carlos also Excel is a well worn comfort blanket for many!!
    17:16:37 From Carlos Mercado : Excel is the worst possible tool for every job in business- but it's always possible.
    17:17:23 From Austin Loveless : It's turing complete I've heard, anything must be possible with it
    17:17:56 From Christian Capdeville : Carlos - not to nitpick, I'm actually interested in your answer to this: Is there a better tool than excel to create custom financial/economic models?
    17:18:10 From Mikiko Bazeley : Google Sheets
    17:18:18 From Mikiko Bazeley : You keep saying Excel
    17:18:24 From Mikiko Bazeley : When you should be saying Google Sheets
    17:18:33 From Rodney Beard : Python is good for economic models
    17:18:36 From Ben Taylor : Find a problem that causes them pain…
    17:18:36 From Russell Willis : @Carlos I'm not so sure?... There's always Power Point!!
    17:18:57 From Ashit Debdas : @carlos yes there is a high chance . my current 99.99% work on excel
    17:19:49 From Austin Loveless : That's wonderful Akshay!
    17:19:57 From Eric Sims : Woohoo! So awesome.
    17:20:01 From Kristen Davis : Congrats to them!
    17:20:08 From Tor Rorvik : Awsome news Akshay....
    17:20:15 From Austin Loveless : They got some great feedback from the group here, glad they used it to make some solid improvements :)
    17:20:39 From Carlos Mercado : Is there a better tool than excel to create custom financial/economic models?

R Shiny, Python Dash. Anything that brings reproducibility and scaled parameter manipulation to the model.
17:20:42 From Thom Ives : Agreed Austin
17:21:13 From Carlos Mercado : Amazon Mechanical Turk
17:21:16 From Carlos Mercado : Fiverr
17:21:25 From Thom Ives : Great Advice Ben!
17:21:27 From Robert Robinson : $10 :-)
17:21:40 From Mikiko Bazeley : I feel like my answer is ways some combination of requests, json_normalize but regex patterns
17:21:53 From Mikiko Bazeley : & regex patterns
17:22:01 From Thom Ives : Scraping does become specialized with each new problem!
17:22:19 From Mikiko Bazeley : I’ll test out the patterns on https://regexr.com/
17:22:47 From Wilson Man : Money solves so many problems lmao
17:22:49 From Austin Loveless : That's a great resource Mikiko!
17:22:51 From Thom Ives : If you do a lot of Beautiful Soup, not a big deal, but if you do it here and there - OUCH!
17:23:20 From Carlos Mercado : Money Solves [.*]
17:23:30 From Emanuel Vassiliadis : Chat parser (in Perl!): https://github.com/vass1138/ArtistsOfDataScience/tree/main/chat
17:23:34 From Wilson Man : I see what you did there, Carlos.
17:23:36 From Mikiko Bazeley : There is value to learning how to parse html though
17:23:41 From Austin Loveless : Thanks for that Emanuel!
17:23:46 From Carlos Mercado : Rvest is great for html parsing
17:23:49 From Thom Ives : Right Mikiko
17:24:04 From Ben Taylor : Darn.. I was trying to log into upwork to take a screenshot of the $10 to prove I wasn’t full of shit but my Zeff.ai google account is no longer valid… no proof that what I said is true.
17:24:31 From Thom Ives : Bummer Ben
17:24:49 From Carlos Mercado : @Wilson thank you
17:24:58 From Thom Ives : I have just a few words on this.
17:25:12 From Austin Loveless : You make it even harder to believe you now Ben. Sounds like an excuse.
17:25:14 From Harpreet Sahota To Thom Ives(privately) : Go for it
17:25:24 From Vin Vashishta : Mikiko nailed it
17:26:07 From Mikiko Bazeley : (This is like my life right now — preparing for data structures & Lagos and going through Standard library)
17:26:18 From Wilson Man : We do? Speak for yourself! ;)
17:26:24 From Mikiko Bazeley : XD
17:26:49 From Akshay Mandke : Thanks for all the inputs in the chat
17:26:55 From Mikiko Bazeley : Yes — every projct
17:27:02 From Carlos Mercado : Just to have my joke explained in the chat. [.*] means "begins with anything, and anything else" i.e. "All".
17:27:09 From Russell Willis : +10 @Mikiko!!
17:27:30 From Kurtis Pykes : @akash - https://slack-files.com/files-pri-safe/T0112TQSZ1Q-F01J75M3PTL/ryan_mitchell_web_scraping_with_python_c__1_.pdf?c=1614986828-bf4bce9f527db4d6 (Web Scraping Book)
17:27:52 From Rodney Beard : Biggest mistake a sorting error while grading an 1800 student Stats class.
17:27:52 From Carlos Mercado : 20 TB of DATA
17:27:56 From Mikiko Bazeley : Serverless ftw
17:28:01 From Kurtis Pykes : @akshay * sorry about spelling
17:28:08 From Carlos Mercado : Cloud, i.e., a computer in Virginia.
17:28:09 From David Knickerbocker : I effectively blew away all of the data in my production database 20 years ago lol
17:28:23 From David Knickerbocker : with no backup....
17:28:30 From Christian Capdeville : Great time as usual everyone - have a great weekend!
17:28:44 From Thom Ives : Ouch Ben! But awesome share dude!
17:28:49 From Mikiko Bazeley : I’ve seen at least 3 companies (that got emails from) where they had some junior entering kill customer data
17:28:58 From Akshay Mandke : @kurtis- no prob. and thanks for the link to the book
17:28:59 From Russell Willis : @ben that data wasn't backed up anywhere else in the data hall?
17:29:18 From Thom Ives : I know I should have a story to top Ben's but I must have blocked it from my memory.
17:29:45 From Ben Taylor : Thom I’m supposed to endorse the same book you did last week and now I’m at a loss on how I can 1-up you… you won that fight
17:30:07 From Carlos Mercado : https://xkcd.com/327/
School calls: "Did you really name your child Robert`);DROP TABLE students;--"
Dad: "Bobby Tables we call him!"
17:30:17 From Thom Ives : I envy you for biggest pain story though. That's pretty good!
17:30:17 From Austin Loveless : I love bobby tables
17:30:33 From Ben Taylor : I have a fail that ended up with me drinking on a plane (raised Mormon, don’t normally drink) and saying F-the-world… hahaha Data Problems...
17:30:43 From Thom Ives : I know you can best me Ben - become one with your inner producer!
17:31:13 From Carlos Mercado : Bubble Sort is not your friend lol.
17:31:14 From Wilson Man : Common theme here seems to be: backing up your data is a good idea. :D
17:31:21 From Carlos Mercado : Version Control your Data!
17:31:21 From Eric Sims : Wait, so Excel isn't a database??
17:31:32 From Mikiko Bazeley : I spent 3 months on a project that got solved in a 1 hr conversation
17:31:47 From Thom Ives : Mikiko - That's awesome!
17:32:04 From Austin Loveless : Oof… lol At least it got solved
17:32:15 From Harpreet Sahota : Tor, vijay, then vin
17:32:53 From Mikiko Bazeley : After spending the three months lol
17:33:03 From Mikiko Bazeley : Like actively working on it lol
17:33:10 From Thom Ives : Great lesson though Mikiko ;-)
17:33:11 From Austin Loveless : "We wasted 100s of hours and destroyed the relationship, but I think it was worth it" - Carlos
17:33:28 From Akshay Mandke : I need a frame that says this ^
17:33:34 From Mikiko Bazeley : Cross-stitched
17:33:38 From Eric Sims : Mark!!
17:33:48 From Mikiko Bazeley : In a cute little embroidery hoop
17:33:50 From Ben Taylor : People that become “Linux bash pros” often over do it… and have oh shit moments…
17:33:53 From Emanuel Vassiliadis : Chat parser relies on text input from Harpreet. If anyone has ideas to automate upload and processing, let me know afterwards.
17:33:56 From Thom Ives : Austin - classic!
17:33:58 From Carlos Mercado : I gotta drop y'all. Thanks though!
17:34:04 From Akshay Mandke : Bye Carlos
17:34:06 From Harpreet Sahota : @Mark we are talking about fails at work, do you want to share
17:34:06 From Austin Loveless : Carlos, can you tweet that so I can get it engraved as a laser tweet? https://shop.lasertweets.co/
17:34:09 From Harpreet Sahota : Peace out Carlos!
17:34:09 From Thom Ives : Ben - I've been there!
17:34:10 From Austin Loveless : Have a good one Carlos!
17:34:14 From Ben Taylor : I’ve had plenty of linux commands where I thought I was a linux bad ass and then executed terrible commands… oh geez
17:34:18 From Mikiko Bazeley : Bye Carlos!
17:34:19 From Vivienne DiFrancesco : I feel like there's a lot of good one liners to keep from these stories
17:34:21 From Mark Freeman : OH DO I!
17:34:23 From Carlos Mercado : Austin whats your twitter, I'll add you on my fake twitter
17:34:34 From Austin Loveless : @ALovelessGuru
17:34:40 From Carlos Mercado : bet
17:34:48 From Austin Loveless : Can't wait
17:34:57 From Mark Freeman : Let me know when to share Harpreet!
17:34:59 From Thom Ives : Take care Carlos!
17:35:07 From Mikiko Bazeley : @ben: I recently went on a zsh spree — locked my computer unzipping 50 gb high-res images for a CV project
17:35:34 From Austin Loveless : ope!
17:35:38 From Mikiko Bazeley : Again - when I was just trying to add a pipe to tree to output the folder structure lol
17:35:42 From Robert Robinson : doh!
17:36:17 From Mikiko Bazeley : Somebody HAS to get fired
17:36:35 From Robert Robinson : Someone has to go to jail. That's the rule. LOL
17:36:42 From Vikram Krishna : LOL
17:36:44 From Mikiko Bazeley : I mean I’m offended as a Japanese person
17:36:52 From Mikiko Bazeley : But also that’s kind of true
17:37:39 From Harpreet Sahota : @Mark - After Vin
17:37:57 From Harpreet Sahota : @Greg after Mark
17:37:59 From Eric Sims : @Vivienne, this question is pure gold.
17:38:14 From Mikiko Bazeley : With that being said, in general, probably best not to make jokes about someone’s ethnicity, gender, or religion
17:38:15 From Austin Loveless : It's pure Bitcoin. We going to the moon
17:38:20 From Eric Sims : LOL
17:38:28 From Mikiko Bazeley : Or sexual orientation
17:38:55 From Austin Loveless : Good call out Mikiko.
17:38:56 From Thom Ives : Ryan Reynolds' Mint Mobile Company has a commercial about a guy that can't remember his Bitcoin pass phrases to his Quarter Billion Dollar Bitcoin account - That guy has us ALL beat!
17:39:31 From Ben Taylor : I watched someone cry once when our 5TB raid failed at the hedge fund. 5 guys standing in a room watching our head of IT cry… good memories
17:39:50 From Mikiko Bazeley : Q.Q
17:40:46 From Thom Ives : Oh wow Ben. That's rough.
17:40:51 From Mikiko Bazeley : Mark- Speaking the story of my life with sales
17:41:33 From Vin Vashishta : Do not share are 3 words the sales team does not understand
17:41:34 From Mikiko Bazeley : (Plot twist: We later find out they were all pictures f the actors who’ve ever played Captain Hook)
17:42:22 From Thom Ives : Vivienne, You've opened a flood gate!
17:42:25 From Eric Sims : Gotta drop off - sad to miss out on the rest of this! Have a good weekend, all!
17:42:32 From Thom Ives : Bye Eric!
17:42:33 From Austin Loveless : Have a good weekend Eric!
17:42:37 From Akshay Mandke : Bye Eric ! Have a good weekend
17:42:41 From Vikram Krishna : Bye Eric
17:42:42 From Vivienne DiFrancesco : So glad I asked this question!
17:43:06 From Rodney Beard : Bye Eric
17:43:20 From Thom Ives : Vivienne, You should be! Many things to help you feel better in the future!
17:43:42 From Ben Taylor : I used to own a rock climbing e-commerce store and a cat pissed on some of our new inventory, one of my co-founders cats. Cat piss corrodes metal, had to throw away $500-1000 product? Seems like a small number but it hurt
17:44:05 From Mikiko Bazeley : Not a single hair then Thom
17:44:11 From Ben Taylor : I’m 37 though
17:44:19 From David Knickerbocker : I'm only 19
17:44:22 From Austin Loveless : I have like 3 wisps I'm catching up
17:44:25 From Ben Taylor : Grayest 30 year old ever
17:44:57 From Austin Loveless : I have an uncle who was gray at 19.. didn't know that could happen
17:44:59 From Rodney Beard : @Daviid Knickerbocker, how does negative age work for making mistakes?
17:45:50 From Robert Robinson : The grass isn't always greener, but sometime it is.
17:46:25 From Vin Vashishta : You say that like its a bad thing
17:46:51 From Robert Robinson : Go for it, Harpreet!
17:46:56 From Mark Freeman : I’m super patchy on my face… One day I’m just going to pull the trigger and get a fake beard :P
17:47:20 From Mikiko Bazeley : Not wrong at all
17:47:43 From Mikiko Bazeley : And it’s one of the defining features of some of the key improvements made in convenient nets archtectures
17:47:53 From Mikiko Bazeley : *ConvNets
17:48:48 From Mikiko Bazeley : Memory optimization can be important
17:48:48 From Austin Loveless : It's work
17:49:03 From Austin Loveless : HE's trying to get the scrapers off his computer and into the cloud
17:49:25 From Thom Ives : Chrome is kind of memory lite I thought.
17:49:27 From Vin Vashishta : Memory, network bandwidth, drive I/O off the top of my head.
17:49:40 From Thom Ives : Agreed Ben
17:49:53 From Thom Ives : Something just sounds off.
17:50:22 From Thom Ives : Experiment
17:50:23 From Austin Loveless : I think he was initially on the free plan and needed to scale up was the response given last time, but no idea what he did from that last week
17:51:14 From Ben Taylor : KISS
17:51:17 From Harpreet Sahota : https://docs.google.com/forms/d/e/1FAIpQLSe0Ysa94r-4Cektb4mQbGi2GHpOPZWmtmzn8elEcj3UjXEL8w/viewform
17:51:18 From Ben Taylor : Works…
17:51:19 From Thom Ives : Ben said experiment - I agree with that
17:51:22 From Harpreet Sahota : https://docs.google.com/forms/d/e/1FAIpQLSe0Ysa94r-4Cektb4mQbGi2GHpOPZWmtmzn8elEcj3UjXEL8w/viewform
17:51:56 From Robert Robinson : Great idea!
17:52:03 From Thom Ives : If Elon can build a tunnel under LA, can't we give awards!
17:52:11 From Akshay Mandke : I filled this one today. So excited
17:52:21 From Vin Vashishta : This is an amazing idea.
17:52:24 From Thom Ives : Get out and VOTE!
17:52:34 From Thom Ives : Filled it out today!
17:52:42 From Robert Robinson : Vote early, vote often. ;-)
17:53:20 From Robert Robinson : Is Dominion in charge of the results?
17:53:37 From Thom Ives : Robert :-D
17:53:40 From Austin Loveless : I'll see if Carlos will post it on his fake twitter.
17:53:52 From Harpreet Sahota : https://docs.google.com/forms/d/e/1FAIpQLSe0Ysa94r-4Cektb4mQbGi2GHpOPZWmtmzn8elEcj3UjXEL8w/viewform
17:54:03 From Ben Taylor : My fake social media account was quoted by USA Today…
17:54:14 From Mikiko Bazeley : Lol fake twiter
17:54:23 From Ben Taylor : Fake me is more famous than real me...
17:54:36 From Wilson Man : Fame is incredibly overrated anyway.
17:54:49 From Mikiko Bazeley : Money is not lol
17:54:55 From Angelo : Good to see you all, have an awesome weekend!
17:54:56 From Kristen Davis To Harpreet Sahota(privately) : Will you drop the link to the Sunday group?
17:54:57 From Ben Taylor : Fame can be engineered… wouldn’t be surprised if the top influencers in the next 5-10 years turn out to not be real, all AI
17:55:00 From Vijay Kumar : Whats the link for sunday zoom
17:55:02 From Thom Ives : Fake us is less inhibited!
17:55:10 From Wilson Man : Money is underrated! haha
17:55:32 From Russell Willis : +1 @Ben There are already fake/AI social media profiles with huge followings and interaction...
17:55:36 From Akshay Mandke : For me personally - Passion, Experimentation, Self- Mentorship, Being Fearless
17:55:39 From Vin Vashishta : Deep fake Vin is more fun at parties
17:56:33 From Jillian Katz : can someone share the info for the Sunday morning calls?
17:56:39 From Mark Freeman : I have a rule that if I give an employer 40hrs, I give myself 10hrs. Used this to learn python and ml!
17:56:50 From Vivienne DiFrancesco : Can someone share the group slack?
17:57:05 From Harpreet Sahota : http://bit.ly/comet-ml-h
17:57:16 From Harpreet Sahota : Sorry, http://bit.ly/comet-ml-oh
17:57:19 From Harpreet Sahota : That’s the Sunday session
17:57:21 From Ben Taylor : Ok… proof that everything I say is true even if it sounds like bullshit:
17:57:22 From Ben Taylor : https://www.usatoday.com/story/tech/2015/08/20/ashley-madison-database-hack-cheating-vatican/32052195/
17:57:29 From Ben Taylor : USA Today article with my fake account
17:57:37 From Ben Taylor : EDITOR'S NOTE: USA TODAY has learned that a source in this story, Klara Jonsson, was using an alias. We still believe the analysis to be correct.
17:57:41 From Harpreet Sahota : Slack: 👋 Let’s move this to Slack! We’ve got 382 folks from the team there already. You can sign up here: https://join.slack.com/t/artofdatascienceloft/shared_invite/zt-dgzn8abm-ge_dKGxrc9Dsuhnly90WTw
17:57:42 From Ben Taylor : USA Today was PISSED!!
17:58:03 From Vivienne DiFrancesco : Thanks, Harpreet.
17:58:23 From Vivienne DiFrancesco : Thanks everyone who shared this week! This group is great. See ya next time!
17:58:42 From Vikram Krishna : Bye Vivienne
17:58:51 From Austin Loveless : Bye Vivienne!
17:59:05 From Thom Ives : I struggle with it too Vin and Ben
18:00:28 From Robert Robinson : Great advice, Vin.
18:01:01 From Mikiko Bazeley : We’re all human — and growing in data science and machine learning is the expression of uniquely human traits, perseverance in the face of seemingly unconquerable odds
18:01:26 From Austin Loveless : This article is hilarious @Ben
18:01:40 From Mikiko Bazeley To Harpreet Sahota(privately) : Greg & Kyle have great stories
18:01:47 From Mikiko Bazeley To Harpreet Sahota(privately) : *Kurtis
18:01:59 From Harpreet Sahota To Mikiko Bazeley(privately) : Thanks, I will bring them on
18:02:09 From Vin Vashishta : What Mikiko said. WOW.
18:02:27 From Mark Freeman : I 100% felt that psychological safety when I broke our product this week. It made such a huge difference. I just had to focus on fixing the problem and sharing with others how to improve our systems. Compared to other jobs where it was full of finger pointing and cover your ass.
18:02:29 From Austin Loveless : @Mikiko I think that many of us come from varied backgrounds and it creates that unique story for what got us excited about it.
18:03:03 From Mikiko Bazeley : @Austin: Absolutely — it was a good first lesson to not take Silicon Valley at first sniff
18:04:25 From Robert Robinson : Goodnight, everyone. Need to run. See you next week. Thanks for the office hours.
18:04:35 From Austin Loveless : Have a good weekend Robert!
18:04:52 From Austin Loveless : Tutorial Hell.. yeah
18:05:05 From Vin Vashishta : Me too.
18:05:15 From Mikiko Bazeley : The Mikiko Effect
18:05:16 From Mikiko Bazeley : Love it
18:05:18 From Mikiko Bazeley : TM
18:05:19 From Austin Loveless : I love it
18:05:44 From Thom Ives : Mikiko's brain needs to be cloned ... or is it her spirit? 🤔
18:05:45 From Tor Rorvik : Another great session.... have a great weekend all..... my pillow is calling me at 1am...... Stay safe..... See you Sunday......
18:05:52 From Austin Loveless : @Thom both for sure
18:05:59 From Austin Loveless : @Tor have a good one!
18:06:05 From Thom Ives : Agreed Austin!
18:06:09 From Thom Ives : Bye Tor!
18:06:10 From Akshay Mandke : Bye Tor
18:06:37 From Thom Ives : Greg is the Top DS student in the World!
18:06:41 From Mikiko Bazeley : Spirit for cloning, brain could be replaced by 30% of Netflix, 30% of 90’s romcom, 40% of the personal self-help selection
18:06:52 From Thom Ives : He get's DS as a businessman too!
18:07:05 From Thom Ives : Mikiko - LOVE IT!
18:07:19 From Austin Loveless : Those are goods percentages though!
18:07:41 From Thom Ives : Love 190% = Priceless!
18:08:04 From Thom Ives : DOH! 90's not 90%
18:08:12 From Thom Ives : OK = 100%
18:08:39 From Austin Loveless : Math on a Friday evening... I feel you @Thom
18:09:03 From Thom Ives : 90's thru me off - wrong units issue ;-)
18:09:07 From Mikiko Bazeley : I love your posts Kurtis and greg
18:09:24 From Austin Loveless : Thank you everyone!
18:09:25 From Mikiko Bazeley : Just great material
18:09:30 From Greg Coquillo : Thank you Mikiko!
18:09:32 From Vikram Krishna : Thank you everyone!!
18:09:36 From Greg Coquillo : You ROCK
18:10:08 From Jaya : Thanks everyone
18:10:26 From Vikram Krishna : Thanks for the session guys. Happy weekend :)
18:10:26 From Akshay Mandke : Have a good weekend everyone
18:10:27 From Austin Loveless : Have a great weekend all!
18:10:46 From Mikiko Bazeley : Bye everyone! Have a great weekend!

Article Comments