The Artists of Data Science

The ONLY self-development podcast for Data Scientists

Chat Transcript from Data Science Office Hours 10


16:54:07 From Carlos Mercado :
16:56:10 From Christian Capdeville : Excellent share Carlos - thank you!
16:57:12 From Kate Strachnyi : hey! I’m with the kids so won’t be joining audio this evening lol. it would be too loud and crazy - good to see you all here :)
16:57:28 From Harpreet Sahota : Thanks for coming Kate! Your presence is felt!
16:57:42 From Thomas Ives : I can relate to that Kate.
16:57:57 From Carlos Mercado : What % of your week is spent reading relevant domain or data science papers / reading newsletters, etc? What newsletters or journals are you following?
16:58:37 From Thomas Ives : Never enough Carlos. Maybe 10%, but my manager and I are thinking I could take more.
16:59:18 From Thomas Ives : Excellent point on projects Gio!
16:59:38 From Ben Taylor : 5% reading audible books since I can multi-task (hike, backcountry ski, workout, bike, etc..). My books are mostly business focused now, less technical.
16:59:51 From Ashen Rana : Currently busy with learning new tools like Snowflake, DBT, and Airflow so I’ve been in the weeds reading documentations and tutorials! What newsletters do you follow, Carlos?
16:59:54 From Carlos Mercado : Just downloaded audible!
16:59:58 From Haseeb Mohammed : i'm overwhelmed with the amount of content available, ive recently unfollowed a bunch of stuff to focus just on databricks for the last month or so.
17:00:06 From Ashen Rana : Finishing up Thinking in systems book also!
17:00:51 From Sasha Prokhorova : I’m reading the Thinking in Systems right now too!
17:01:49 From Harpreet Sahota : If you guys have a question for the crew here type it out and hold your place in “line”!
17:02:34 From Eric Sims : Question: What is unit testing? What is it used for?
17:02:57 From Carlos Mercado : My newsletters are:

Andriy Burkov's AI Newsletter
Emerging Tech Brew
Data Elixir (just joined this one)
The Forecast

17:03:09 From Carlos Mercado : It's a lot but its both technical, pop-AI, and Sales.
17:05:44 From Ashen Rana : Nice list Carlos! Will check em out
17:07:01 From Ashen Rana : Could you please speak to “a significant shortage of business domain experts in the data science and analytics space”? What’s your observation? Prediction?
17:07:25 From Ashen Rana : ^ quote from Jason Krantz’s post
17:08:01 From Kate Strachnyi : not sure if this is the right forum to do this but I had a request today from a contact that’s looking to hire -Bank of America account team for Pega. The bank is looking to do a direct hire of a Data Scientist to help develop insights for Consumer Bank clients.
17:08:21 From Carlos Mercado : does it have a citizenship or location requirement?
17:09:28 From Haseeb Mohammed : re: recruiters
17:10:05 From Carlos Mercado : @Ashen:
A lot of people are learning data science as:

problem -> Python -> Deep Learning -> Results -> look at me.

Instead of problem -> context -> hypothesis -> exploration -> discussion -> model -> does model pass common sense tests -> identify if this problem is solved or if it is now a monitoring problem that needs continuous development / an API / cloud hosting of a model / delivery of insights, etc.
17:11:00 From Ashen Rana : That’s a good way of breaking it down, Carlos - thanks!
17:12:24 From Mark Freeman : What’s your approach for V1s of implementing data science solutions? In my role, I’m the first data scientist in the company as they are now ready to work towards personalizing their product. Thus I’m implementing a lot of V1s at my org.
17:15:51 From Ben Taylor : The more you MAKE them care about the tools and details you used to solve the problem… the MORE they want to fire you :)
17:16:12 From Karan Ambasht : Lol !!
17:16:29 From Christian Capdeville : lol!
17:16:30 From Karan Ambasht : Thanks Carlos, Monica
17:17:33 From Carlos Mercado : @ Mark
figure out really quickly if what you're programming is going to be needed in 6 months; if the vast majority of your work right now is going to get dumped; then just go for speed and iteration.

Make as many models and charts and outputs and features that might be useful; be in constant communication with the team; and avoid overcommitting to useless things. Like Ben said earlier; get 1-month timelines not 6-month moonshots.
17:17:46 From Sasha Prokhorova : I’ve been wanting to learn about the best time managing practices as well! Multi-tasking vs. uni-tasking?
17:18:22 From Haseeb Mohammed : @Ben can I quote you
17:18:41 From Ben Taylor : @haseeb Always
17:19:04 From Ashen Rana : The Pomodoro technique has been effective for me Sasha! Work on tasks in a 25 minutes block, without distractions
17:19:07 From Carlos Mercado : time management
(1) don't beat yourself up. If you drop something, oh well, you dropped it.
(2) consistent 15mins beat 2-hr random sprints.
(3) leverage "flow" state. if you're not feeling like doing something, go do something else for a bit. I code in sprints; I can't just sit there and bang my head for hours.
17:19:10 From Manna Sirak : So it makes sense that ideally you'd have a focused portfolio/projects to present to interviewers. Any thoughts on how to structure a DS/analytics portfolio if you don't know exactly what industry you want to be in?
17:19:42 From Haseeb Mohammed : Carlos> (1) don't beat yourself up. If you drop something, oh well, you dropped it.

17:19:46 From Haseeb Mohammed : do what you said you would do
17:19:47 From Carlos Mercado : Google GitHub project outlines; there are a lot of optimized file path structures for certain tasks (e.g. what gitignores to use, file naming, etc.)
17:19:54 From Haseeb Mohammed : so if you find yourself dropping the ball, figure out how to say no
17:19:55 From Carlos Mercado : Ok ok ok - yes
17:20:09 From Ben Taylor : @manna I would fight ‘sameness’, everyone’s resumes look the same. Pick projects you are passionate about, that wake you up early on the weekend because you’re so excited to invest in those projects. Those are the best. I look for passion.
17:20:20 From Christian Capdeville : Rewrite your priority list each day or week.
17:20:34 From Christian Capdeville : It will change - number 2 this week may be number 6 next week
17:20:35 From Carlos Mercado : if you told someone else you'd do something, do it. But if you don't study one night and then you just hate yourself and quit studying; then that was a waste.
17:20:54 From Haseeb Mohammed : agreed
17:21:38 From Carlos Mercado : @Haseeb I am trying to say this:
17:22:06 From Haseeb Mohammed : that was a great response from the govinator
17:22:16 From Kate Strachnyi : best time management tip is saying no to anything that doesn’t support the main goals
17:22:49 From Carlos Mercado : My mentees will waste 2 weeks beating themselves up for not getting something done. It's like, if someone steals $10 from you, do you delete your bank account??
17:23:49 From Carlos Mercado :
17:24:03 From Manna Sirak : @Ben Thank you!
17:24:25 From Ben Taylor : You get better at what your practice too. Things you think you are weak at (if you want them to be strengths) they will get better over time. I used to hate moderating a panel, I sucked at it, now I’m fine with it.
17:24:48 From Giovanna Galleno : great thoughts @Ben!
17:25:34 From Ben Taylor : The power of procrastination…. When deadlines come… shit gets done. You will fill the time you give for the projects you have, don’t give yourself too much time.
17:25:47 From Thomas Ives : Joe!
17:25:49 From Ashen Rana : ^ Being able to say no to people (politely) is a big hurdle at first but so rewarding in the long run
17:25:55 From Joe Reis : Howdy!
17:26:03 From Haseeb Mohammed : hi joe!
17:26:06 From Ben Taylor : HEY Joe!
17:26:12 From Eric Sims : I'm pressure prompted, so the power of procrastination is real!
17:26:22 From Carlos Mercado : @ Ben - 100% agree; I can do a week's worth of work in 1 14-hour sprint if I don't get a choice.
17:26:24 From Ashen Rana : Procrastination got me through college lol @Ben
17:26:43 From Ben Taylor : Take a f*cking nap if you need it!!! I’ve spent 2 hrs on a damn bug that took me 4 min after a nap
17:27:11 From Haseeb Mohammed : told my teammate this week to take a nap
17:27:14 From Haseeb Mohammed : he came back after 2 hours
17:27:16 From Haseeb Mohammed : and crushed it
17:27:21 From Haseeb Mohammed : WFH >
17:27:23 From Joe Reis : Naps are nice
17:27:43 From Giovanna Galleno : Agree @Joe!
17:27:54 From Sasha Prokhorova : Siesta is the king.
17:27:57 From Thomas Ives : Computer screens are the biggest hindrance to next level troubleshooting sometimes in my experience.
17:28:11 From Thomas Ives : AND lack of sleep.
17:30:48 From Ben Taylor : Carlos time limit yourself on that type of effort/project as well, that could easily consume weeks or months of “R&D”. What is the 2 day effort that could tell a story that a SME could get. You’re hunting for intuition
17:31:40 From Carlos Mercado : I think I needed to hear "2-day effort", we keep doing stand-ups and I have updates and new visuals, but it isn't translating because it doesn't have that KPI. That's the missing piece.
17:32:12 From Carlos Mercado : Thanks!
17:33:24 From Harpreet Sahota : Ted Petrou, Dunder Data, Build a Data Analysis Library from Scratch
17:33:45 From Manna Sirak : Have to head out - thanks all!
17:33:53 From Thomas Ives : Thanks Harpreet
17:33:53 From Eric Sims : See you Manna!
17:34:00 From Thomas Ives : Bye Manna
17:34:06 From Haseeb Mohammed : take care folks, i'm out as well
17:34:11 From Giovanna Galleno : Thanks for joining Manna! All the best! :D
17:34:47 From Giovanna Galleno : Thanks! Take care @Haseeb :D
17:35:02 From Carlos Mercado :
Great book on "don't scale until it hurts"; free too.
17:35:20 From Dave Langer : Go, Ben!
17:35:34 From Dave Langer : The Grim Reaper argument, love it!
17:36:43 From Nicholas Urban : Gotta run, thank you all!
17:36:58 From Karan Ambasht : Thanks everyone, got to head out !!
17:37:16 From Ben Taylor : Pinch your time, not your pennies
17:37:27 From Suraj Bondugula : Got to go. This was really fun and informative. Thanks, Harpreet for organizing this. Helps me a lot in knowing things. Bye!
17:37:56 From Ben Taylor : VIM shortcuts
17:38:01 From Joe Reis : ^ true
17:38:17 From Joe Reis : Think outside the algorithm
17:40:33 From Joe Reis : I’m actually a big proponent of talking to people
17:40:34 From Eric Sims : Genchi genbutsu! Go to where the work gets done.
17:40:39 From Ben Taylor : I’ve sat down with someone working the process (an underwriter) and taken notes for 2hrs… it was SO insightful… more than any VP or director could have communicated.
17:40:46 From Joe Reis : yep
17:40:52 From Ben Taylor : Who has worked the process the most you are trying to automate, talk to them
17:41:05 From Thomas Ives : ^ great examples.
17:41:12 From Thomas Ives : Great points Gio
17:41:19 From Dave Langer : Folks in the trenches usually know how the process REALLY works.
17:41:38 From Ben Taylor : I’m an expert “Googler”
17:41:39 From Carlos Mercado : but maybe don't tell them your plan is to automate them out of their job without a plan for them. We have so much change management problems with my clients who have employees terrified of being fired by effective consulting.
17:42:04 From Kate Strachnyi : great conversation all! have to go now but glad I joined :)
17:42:12 From Joe Reis : Later
17:42:15 From Ben Taylor : @Carlos I’ve found that people don’t typically lose their jobs, their jobs become less boring. They can finally be creative with a process that has been stale for way too long
17:42:17 From Sasha Prokhorova : What are your thoughts on unpaid internships? Do you think they bring value for the newcomers, or do you find them to be unethical practices? Or is there no “cookie cutter” answer?
17:42:18 From Dave Langer : Wisdom, Carlos. Long ago I worked for a P&C insurance company and I sat with Claims people to do exactly that. 😲
17:42:22 From Carlos Mercado : That's 100% true
17:42:38 From Eric Sims : See you later, Kate!
17:42:39 From Ashen Rana : That Dave. Over time those folks probably have figured out shortcuts to be more efficient
17:42:56 From Carlos Mercado : But people, especially with my (government) clients, enjoy that their jobs for 5,10,15+ years has been mentally automated. They can do it in their sleep. Boring is a feature sometimes.
17:43:02 From Joe Reis : “What would you do here” - The Bobs
17:43:09 From Thomas Ives : That's a hard one Sasha. It seems to depend on many things.
17:43:22 From Ben Taylor : Office space… sooooo good!
17:43:59 From Carlos Mercado : unpaid internships solidify societal inequalities. That is a reason to not have them. Although I would support education/institution based co-ops where a 3rd party funds an internship (school, tuition, government, etc.)
17:44:18 From Ashen Rana : Office Space was funnier before I worked full time lol. It’s my nightmare now
17:44:29 From Joe Reis : It’s actually a documentary
17:44:37 From Carlos Mercado : especially for small businesses who NEED analytics support and can't afford real help and are getting drowned by big competition.
17:44:52 From Ben Taylor : I don’t like Silicon Valley…. I relate WAAAYY to much to those stupid episodes...
17:45:08 From Ashit Debdas : my work background not related to IT and data stuff, how can I do transaction ?
17:45:12 From Ben Taylor : Like picking the name of a company… why we changed our business name from Ziff to Zeff……
17:45:15 From Joe Reis : Season 2 of SV was a documentary when we were raising our Series B. It was sad
17:47:07 From Carlos Mercado : Things data scientists need to know:

Reproducible code
how to name files and variables
dependency mgmt. (i.e. package versions)
Git for version control
17:47:12 From Thomas Ives To Harpreet Sahota(privately) : You've earned this and more. Treat it like the new normal

Article Comments