Starting to see more articles like this. The field...
# datasciencet
thomasnield
10/27/2018, 10:31 PMStarting to see more articles like this. The field of "data science" is increasingly going to draw questions on its definition, and perhaps cause abandonment of the term.
https://www.datasciencecentral.com/profiles/blogs/why-do-people-with-no-experience-want-to-become-data-scientists
v
ValV
11/09/2018, 6:01 PMEveryone wants a big salary and a prestige job. If a man has no strong skills in a particular field, then he needs such a job twice as much (perhaps, psychology / essential needs). 👈 As for the article. I think it's just like why people want to be actors, artists, managers, etc. It's kinda some hard-coded (genetically) desire of superiority (if not salary, then prestige). And we can't solve this problem easily. I think, even if the title data scientist will die (as proclaimed by the author) -- another will appear, perhaps, with the same level of obscurity and hype. The words of the author are full of irritation, obviously because of the newcomers, that pretend the same salary/prestige as he (without 20 yrs experience)
ValV
11/09/2018, 6:03 PMLyrics 😁
🎵 1
🎤 1
ValV
11/09/2018, 6:06 PMCan I call myself a data scientist, if I had won, let's say, two or three Kaggle competitions (in the top 3, but not the first)? 😎
t
thomasnield
11/09/2018, 7:18 PM@ValV Well I think you need to get in first place in 2-3 Kaggle competitions before you're given a peace prize, for predicting the probability an individual will die on the Titanic.
thomasnield
11/09/2018, 7:59 PMI have found this one quite interesting. I'm not doing the contest but I forked it for research later.
https://medium.com/crowdai/can-you-make-swiss-trains-even-more-punctual-ec9aa73d6e35
⭐ 1
v
ValV
11/09/2018, 8:11 PMCool, that's much better application, than _AlphaPilot_'s ('cause we know why do they need AI drones)
ValV
11/09/2018, 8:13 PMI didn't know about crowdAI, please share more 😊
d
dalexander
11/09/2018, 8:23 PMThat looks pretty cool.
t
thomasnield
11/10/2018, 4:57 AMYeah speaking of, I actually prototyped an algorithm for work tonight that executes on-time performance planning. It mitigates delays by optimizing planned times fitted to an on-time performance goal, like 80%.
thomasnield
11/10/2018, 4:59 AMI wish I could share it, but I probably shouldn't for industry competitive reasons. I will say that discrete optimization keeps surprising me in how powerful it is, and it's astounding how often it is overlooked.
v
ValV
11/10/2018, 5:19 AMHmm, interesting... But some question arose about optimizing timetables. E.g. with that example on Swiss railways -- if there a crash happens on a branch -- it should cause more expenses against the case with non-optimized timetables? I'm not taking into account that a company will gain more profit from optimized timetables to recover from the crash 😏 But in a small time window it should cause more expense
ValV
11/10/2018, 5:23 AMOr are there alarm systems that detect crash quickly and rebuild routes?
t
thomasnield
11/10/2018, 5:28 AM@ValV the thing with optimizing timetables is you aggregate all the mess statistically and not get that granular ( hopefully crashes don't happen that often!). If you are late 70% of the time (regardless of the reason), you can try to pinpoint specific factors. But you are more likely going to find ways to pad timings without sacrificing utilization too much, and bring it up to something more acceptable like 80% or 90%.
5 Views