South African COVID-19 Vulnerability Map by #ZindiWeekendz
Helping Africa
$300 USD
Completed (over 6 years ago)
Prediction
318 joined
177 active
Start
03 Apr 20
Close
05 Apr 20
Reveal
05 Apr 20
'%3e%3ccircle r='20' fill='%23008'/%3e%3ccircle r='17.5' fill='%23fff'/%3e%3ccircle r='3.5' fill='%23008'/%3e%3cg id='d'%3e%3cg id='c'%3e%3cg id='b'%3e%3cg id='a' fill='%23008'%3e%3ccircle r='.875' transform='rotate(7.5 -8.75 133.5)'/%3e%3cpath d='M0 17.5.6 7 0 2l-.6 5L0 17.5z'/%3e%3c/g%3e%3cuse xlink:href='%23a' transform='rotate(15)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='rotate(30)'/%3e%3c/g%3e%3cuse xlink:href='%23c' transform='rotate(60)'/%3e%3c/g%3e%3cuse xlink:href='%23d' transform='rotate(120)'/%3e%3cuse xlink:href='%23d' transform='rotate(-120)'/%3e%3c/g%3e%3c/svg%3e)
Is this a data science competition or a marathon ?
Help · 4 Apr 2020, 15:57 · edited 7 minutes later · 9
Hello everyone,
I don't participate regularly to zindi competitions but I decided to join today and I m seeing people submitting more than 200 times in 24 hours :p
I also saw people saying that train and test are from completely different distributions. What is your strategy when submitting ? or you just try one feature and submit => see results , if score drops : remove that feature and try another one. If that's the case this is not data science at all.
I mean to win here, you don't need to be an expert in the domain, you only need to have a lot time to spend ( correct me I m wrong ) I don't want to minimize people in the platform I m sure there are many top data scientists here.
Enjoy the rest of the contest
I think Zindi should have limited the number of submissions to 10-20 a day, the more number of submissions you make the better chances you have at winning, its simple maths, it increases the probability of winning. Yes, I agree making so many submissions is not data science, its hit and trial. Enjoy everyone.
correct mishra! you are always on point
This is not data science but competitive data science is. The end result is to score the highest on the leaderboard. How you get there, as long as it's within the competition rules, is for you to decide.
thanks so much, we all have equal Chance of 200 submissions its within the Rule, it's 3 days Hackathon not 2/3 month it's just to learn and catch fun.
I don't agree when you say it's just to learn. People learns how to overfit public LB without making generalizable models which is an important skill as a data scientist
And this I believe is a serious issue, with regards to solutions to be deployed for potential global use...... Overfitting
I think it should be even much better if we are only given a public lb of about 10-15 %.The rest be on private lb.A suggestion though
I totally agree, Another suggestion is to reduce submissions for this weekend hackathons. Zindi should please look into this.
This is data science; Concerning this; 'if score drops : remove that feature and try another one' Kindly read up on various feature selection, solving a linear equation, trinomial and its mathematical concept; There is more response to your post.. so the best thing is to reduce the submission to 10 per day... Be good man