@Carl Love 

Sorry, in France we don't learn much statistitics during high school, some binomial, some Gauss law ..

i will obviously test once I get the datas different laws, including the continous Markov chain that you describe.

But more important, is to find if it exists a trend to increase ou decrease during these 6 years.

Last year, our head department said that we could stop this task, as (that's what he thought, he's retired now) we did such a good job in another place that near zero consultations per day were made in urgencies.

Actually, I see (and our new head department too) that we're rarely under 2-3 consultations per day (last 15 days, that was once only under 5 consultations asked, and twice 7 consultations)

Thus, what is your advice about the more powerful way (meaning the second order risk) to find if a trend exist ? 6x325 observations are something, but on the other hand a mean of 2 consultations per day will probably don't help..

Firendly yours,

Milos

@Carl Love 

I'm not sure to fully understand your reply, as I always believed that Markov chains were really hard to compute.

But obviously I will follow your answer, as it sounds as a perfect match, and if it is trivial to do with Maple..

A lot of thx for your answer.

Friendly yours,

Milos

@DSkoog 

Sorry to bother you with silly questions, I found how to do :

gusen_d := DataFrame(gusen);

I'm really stupid.

Thanks again for your precious help.

Friendly yours,

Milos

@DSkoog 

Thanks again, DataFrame sounds exactly as a match with my Excel file.

I got my matrix with :

gusen := Import("C:\\documents nega\\fiche gusen octobre novembre 41 consommation coke fours.xlsm");

Does exist a method to create and import directly this Excel file in a DataFrame ?

This file has 28 lines, but I will soon have another file with 1800 lines to study (each line with the number or rather rare occurences, likely following a Poisson law) so it could be somewhat useful to import directly such a file..

Thanks for your help again, friendly yours.

Milos

@Carl Love 

Many thanks for your answers. I think that Excel can convert dates to a float.

Thus, I have now to convert the columns to vectors and do that each element of a vector is apparied do the another. Thx to you, I shoud be able to do that .. :-)

@Carl Love 

Many thanks for your answer.

I want to extract 2 columns (the matrix is rectangular, 28x7), especially the column 4 and 7, optionnally the first (dates, with here an error converting dates to floating numbers), adding that the first row contains the data names, thus the real datas are in the column 4 and 7, starting from raw 2.

Does exist a mean to extract theses datas (4 and 7, perhaps 1 converted to floating numbers with Excel) excepting the raw 1, or indicating that it is the name of each data columns ?

Thx again for your help,

Milos