In the future, you will need to understand deep statistical concepts by reading technical articles. As a training for that, it is interesting to start from here. Also, as we have limited time, this is a way to have some self guided learning to understand everything better and have a wider knowledge.
This week you will find some questions here that you will need to answer by documentating yourself. So you will do a different PR for each question (you are meant to answer the questions in different days). Don't hesitate to write as many text as you need and push images if you need them.
Remember for this lab: there is a right answer. But there is no perfect way to explain it (except for in a mathematical way, but this is another story).
You know both concepts but, is there a difference? Are they synonims? Start investigating.
As a good reference (once you have lookedfor some information) you have this article.
Oh no dear, they are absolutely not. In statistics and probability analysis, the expected value is calculated by multiplying each of the possible outcomes by the likelihood each outcome will occur and then summing all of those values. The mean on the otherhand represents the average of a specific sample.
--------------------------------------------------------------------------------------------------------
We have told you that a lot of scientifical investigations are based on p-values. The last week, Nature magazine published an article regarding the problem. Start digging on it!
Don't hesitate to use more articles if you want to :)
A/B testing is a widely used tool to understand differences between two samples. It is a way to measure the impact of something we did:
- A marketing campaign.
- A new feature in our application.
- A new design in our application.
- A different flow in the User Experience flow.
To do this, is very important first to design our experiment.
- We need to know how we are measuring the impact. If people has the behaviour we want with this new implementation.
- We choose a control group (people who doesn't have/see the new change) and the group which will see the new change.
- We think about how much data do we need.
- We measure the difference between them.
One example: Our application has a lot of mini-games. We want people to reach the games that we think are the best but the behaviour is not the expected, they don't reach them.
So we call a designer and after a lot of work he shows us a new design for our application: we will add a botton specific for that kinf of games inviting the users to click on it:
Click here to discover cool games!
We think it will work but can we be sure? So instead of implementing this new botton for all users, we implement it for 10% and we compare the results with the users that didn't have it. Is there a significant difference? Is our botton working?
Read more about A/B testing with a couple of examples:
Another example about Netflix here
You need to submit a markdown file with the answers to the questions above. You can create a new .md file or directly edit the README.md.
Upon completion, add your deliverables to git. Then commit git and push your branch to the remote.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent