New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Docs example using to_dataframe() #436

Open
braaannigan opened this Issue Feb 15, 2018 · 8 comments

Comments

Projects
None yet
3 participants
@braaannigan
Contributor

braaannigan commented Feb 15, 2018

I think that the to_dataframe() method in #425 will be the easiest way to access samples once it's merged. I want to add some documentation showing how it works. I was planning on doing a simple linear regression example - does anyone have an objection or any comments before I start?

@ahartikainen

This comment has been minimized.

Show comment
Hide comment
@ahartikainen

ahartikainen Feb 15, 2018

Collaborator

Sounds good.

Are you considering to manipulate the dataframe after the export? (E.g. adding x as index)

Collaborator

ahartikainen commented Feb 15, 2018

Sounds good.

Are you considering to manipulate the dataframe after the export? (E.g. adding x as index)

@ahartikainen

This comment has been minimized.

Show comment
Hide comment
@ahartikainen

ahartikainen Mar 7, 2018

Collaborator

Fyi notice #444 and issue #443

Collaborator

ahartikainen commented Mar 7, 2018

Fyi notice #444 and issue #443

@braaannigan

This comment has been minimized.

Show comment
Hide comment
@braaannigan

braaannigan Mar 12, 2018

Contributor

Issues noted @ahartikainen.

I've got some time to work on this now.
Question for @ahartikainen @ariddell: rather than me making up some random examples, do we want to source examples (and small datasets) from somewhere e.g. the McElreath book?

Contributor

braaannigan commented Mar 12, 2018

Issues noted @ahartikainen.

I've got some time to work on this now.
Question for @ahartikainen @ariddell: rather than me making up some random examples, do we want to source examples (and small datasets) from somewhere e.g. the McElreath book?

@braaannigan

This comment has been minimized.

Show comment
Hide comment
@braaannigan

braaannigan Mar 12, 2018

Contributor

Thinking about this a bit more, I'd prefer just to use simulated data as it avoids issues with keeping links up-to-date etc.

Contributor

braaannigan commented Mar 12, 2018

Thinking about this a bit more, I'd prefer just to use simulated data as it avoids issues with keeping links up-to-date etc.

@ahartikainen

This comment has been minimized.

Show comment
Hide comment
@ahartikainen

ahartikainen Mar 12, 2018

Collaborator

I don't see a problem if we use some external data.

One reason for this is that I always find it really confusing when I see examples of using simulated data (as their only data source). It is missing the step where one is trying to guess the real model.

In my mind an example of fitting a model goes like this (option 1):

Real data --> plot it --> think about the model which created the data --> create a model and sample from it --> compare simulated data and real data (this is the step where one usually notices that things are not iid) --> update model (loop until satisfied) --> fit parameters for simulated model --> check it works --> fit parameters for real data --> describe findings and update model if needed (loop).

And normally they are (option 2):

fit model to data --> print fit --> stop.

edit. this is just me ranting. If you use simulated data, I guess it should be done before anything else

Collaborator

ahartikainen commented Mar 12, 2018

I don't see a problem if we use some external data.

One reason for this is that I always find it really confusing when I see examples of using simulated data (as their only data source). It is missing the step where one is trying to guess the real model.

In my mind an example of fitting a model goes like this (option 1):

Real data --> plot it --> think about the model which created the data --> create a model and sample from it --> compare simulated data and real data (this is the step where one usually notices that things are not iid) --> update model (loop until satisfied) --> fit parameters for simulated model --> check it works --> fit parameters for real data --> describe findings and update model if needed (loop).

And normally they are (option 2):

fit model to data --> print fit --> stop.

edit. this is just me ranting. If you use simulated data, I guess it should be done before anything else

@tommylees112

This comment has been minimized.

Show comment
Hide comment
@tommylees112

tommylees112 Aug 13, 2018

Has this been merged to the current working version or only in the dev version? Great functionality thanks guys!

tommylees112 commented Aug 13, 2018

Has this been merged to the current working version or only in the dev version? Great functionality thanks guys!

@braaannigan

This comment has been minimized.

Show comment
Hide comment
@braaannigan

braaannigan Aug 13, 2018

Contributor

Hi - it's in neither unfortunately, life intervened when I had barely got started. I don't expect to be able to work on anything before next summer due to continued life interventions!

Contributor

braaannigan commented Aug 13, 2018

Hi - it's in neither unfortunately, life intervened when I had barely got started. I don't expect to be able to work on anything before next summer due to continued life interventions!

@ahartikainen

This comment has been minimized.

Show comment
Hide comment
@ahartikainen

ahartikainen Aug 13, 2018

Collaborator

Edit. Sorry I misread the issue.

Is currently on the dev and next version 2.18 will be released in a day or two.

Collaborator

ahartikainen commented Aug 13, 2018

Edit. Sorry I misread the issue.

Is currently on the dev and next version 2.18 will be released in a day or two.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment