Gathering statistics

[atmtaatmta]

Hello everyone.

I am going to write a program on this weekend, which will model applications arriving in pool and CIC draws to predict the outcome.
There's not really enough data to create a really accurate statistical model, but I think I can manage to create a crude but usable model using spreadsheet and info about two draws already done.
Unfortunately I lack information about number of people in the pool and the number of new applications submitted for each day (it surely changes over time).

The number of profile seems to be a good hint. Since it is E followed with numbers, I think that stands for Express Entry only. I also think, that first ~200000 were used for fake accounts during testing. That way, knowing the rough number profiles and dates they were submitted, it is possible to find out the speed of pool filling and thus create a probability model to see the scores being invited in the vicinity and the marginal scores for ITAs when when they stabilize.

After I am done modelling I will post the results here of course. And if I have enough time to implement a decent interface, I will post the modelling program itself, so you can all play with numbers and see the results for yourselves.

So I need you to post your number (you can omit last 2 or 3 digits, since it can be considered to be a personal information and I won't need it anyway) and the date, when the profile was submitted.

I'll be the first:
E0002199**
January-2-2015 03:26:19 AM EST

 

[atmtaatmta]

it might not be continuous, it might include date or time info and what not, you don't know what number they started with, etc.).

There have already been topics, where people posted their numbers with dates.
It surely was continous and it is unlikely, that there is date included.
Alas, there was too little data to make precise statistical conclusions. That's why I need as much info, as community can provide.

I'm also skeptical about the value of an elaborate statistical model if you make wild guesses about pool size or growth rate. The best information will come from the upcoming draws.

Ok, look.
We have the distrubution of scores thanks to spreadsheet. There are about 3% of all applicants from the pool. That means that this distribution is quite accurate.
Suppose we know how many people submitted profiles every day. This way we can model the probability process of new applicants getting into the pool over time.
Since the numbers are large the result will be close to reality.

And also we now have two draws to test the model. If real ITA scores will be close enough to predicted, that will mean, that we can predict the nearest draws. And if there won't be any turbulances in application submitting process, given the target CIC annual target numbers of invitations, we can predict the minimal score, which will have ITAs isuued.

At last, there is no harm in trying. If model would be too anaccurate, that would mean, that distribution in pool differs from the one in spreadsheet. That would also be rather valuable information.

 

[atmtaatmta]

I understand statistical modelling and I know exactly what you want to do, but like you say it will all depend on your assumptions (for some of which you do not have enough information for):

- size of pool
- growth rate
- number of job offers / provincial nominations
- actual target

That's why I created this topic: to find out size of pool and growth rate.
I don't care about number of jos and pns, because it doesn't matter in fact as we have score distribution thanks to spreadsheet.
CIC has already posted actual target for 2105 back in 2014.
So, while the information is still not complete, that's enough to make assumptions.

Your assumption that the first 200,000 tracking numbers are void is just a guess. It could be 100,000 or none. When you consider such huge variations, any model you can come up with will just produce meaningless numbers.

It is not. Well, it is a guess, but not wild. There were 20 000 - 40 000 account in the pool at the end of January. As you can see, my number starts with 219 *** and it was posted early on January, 2. Profile submitting became operational just 15 hours before that. I really don't think, that more, than 10 000 people registered before me.
There are also people, who managed to submit their profile on Januray, 1. If we knew their application numbers, it would be more clear, what the starting numbers are.

 

[fl_pie]

That's why I created this topic: to find out size of pool and growth rate.
I don't care about number of jos and pns, because it doesn't matter in fact as we have score distribution thanks to spreadsheet.
CIC has already posted actual target for 2105 back in 2014.
So, while the information is still not complete, that's enough to make assumptions.
It is not. Well, it is a guess, but not wild. There were 20 000 - 40 000 account in the pool at the end of January. As you can see, my number starts with 219 *** and it was posted early on January, 2. Profile submitting became operational just 15 hours before that. I really don't think, that more, than 10 000 people registered before me.
There are also people, who managed to submit their profile on Januray, 1. If we knew their application numbers, it would be more clear, what the starting numbers are.

People creating profiles 3-4 hours after the system went online got 217*** or 218*** numbers.

 

[DesireCA]

I'll be the first:
E0002199**
January-2-2015 03:26:19 AM EST

E0002600**
Febuary 11, 2015 (~1200 Hrs GMT+5)

Lets hope for the best

 

[atmtaatmta]

E0002600**
Febuary 11, 2015 (~1200 Hrs GMT+5)

Lets hope for the best

Thanks, that will be handy. Now we know, that are about 42 000 profiles.
Still need more data, though.

 

[AAA2133]

E0002200xx

Submitted: January-2-2015 04:29:13 AM EST

Not sure when it was created though!

 

[fl_pie]

Thanks, that will be handy. Now we know, that are about 42 000 profiles.
Still need more data, though.

Also reposting from a different thread: I know that someone who registered on January 28 received E00025XXXX number. So yes, it does seem like the number is continuous.

 

[fl_pie]

Maybe somebody who has a Twitter account could ask the CIC if they are ever planning to disclose how many applicants are in the pool.

 

[atmtaatmta]

Bump. Need more data.

 

[fl_pie]

Btw, does anybody have an ID which starts with E000216XXX or a smaller number?

 

[nads1]

E000217XXX not smaller but close to what you posted

 

[fl_pie]

E000217XXX not smaller but close to what you posted

Thanks. I'm still curious if there are any lower numbers. My guess is that numbering is continuous, but that it started from some big number, maybe E000215000 or something like that, so I'm trying to figure out what the starting number was.

 

[atmtaatmta]

Found some data here: http://www.canadavisa.com/canada-immigration-discussion-board/-t268439.0.html
E000244XXX, January 23 - 2015
January 28 E00025XXXX
I know there's more, I've seen it, just can't find.

 

[atmtaatmta]

Are you guys sure whether or not an eligible profile receives a reference number?

E.g. the number could be created when someone first creates a profile and be discarded if it turns out to be ineligible.

Just a thought...

I am perfectly sure, that ineligible profiles also have their numbers. There are crude means to calculate the percentage of ineligible profiles though.