You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have some questions about the structure of these dialogues. I have a minor question about One Variable first, and then particularly questions about the 2 variables. They need to be answered, before we consider the structure of the 3-variable dialogues.
One variable:
a) I am particularly concerned with the Describe > One Variable > Summary.
b) We currently have 2 options, namely the default summary function - which I like. Plus our improving Customised option that is starting to look really good. (It is roughly a special case of Specific > Tables, when there are no classifying factors.
c) There is a tidymodels alternative to summary called skim that will probably be useful as an alternative to the original summary, rather than a replacement?
d) We are going to be examining different ways of adding a group_by to R-Instat. I suggest there will be a variety of ways that we want to add group_by. I suggest we consider adding it here as an option for skim. That's because skim should produce tidy output (I assume). So including skim by a factor - i.e. thinking of it as an extension of one variable could be worth considering. This could be here, or as an option for 2 variables, see below.
Now the big one, namely 2 variables.
Currently, in 2 variables, (graph and summary) we are looking at one situation only. This is where we examine multiple variables of the same type, by a single variable. I would like to add further options - presumably through our usual radio buttons.
a) One reason that the ggpairs (that we examine in the correlations dialogue) does cope with factors as well as numeric. We are missing a great low-hanging fruit ignoring it!
b) So I suggest another option (button?) is perhaps called Pairs under 2 variable graphs, which just has a multiple receiver. And the data can at least be both numeric and factor (and we examine what happens with date, logical, character just to be sure. There is a wide range of options and this should be an exciting addition.
c) On the Summarise there are (at least) 2 interesting options. The first is for numeric and would be to put correlations (of pairs) as an option under 2 variables - maybe as well as multivariate. For simplicity (and because there are graphs as well as numerical summaries) I wonder about keeping it as a separate dialogue. There is another possibility here, namely to add (or change) to using the corrr package.
d) Then should we investigate the skim (tidymodels) with a by groups as an additional option for the 2 variables summarise. Then the multiple receiver can have data of mixed types and the second has to be (or is treated as) a group by variable? (If so, then one option for 3 variables will be skim with 2 factors as by variables.)
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
I have some questions about the structure of these dialogues. I have a minor question about One Variable first, and then particularly questions about the 2 variables. They need to be answered, before we consider the structure of the 3-variable dialogues.
One variable:
a) I am particularly concerned with the Describe > One Variable > Summary.
b) We currently have 2 options, namely the default summary function - which I like. Plus our improving Customised option that is starting to look really good. (It is roughly a special case of Specific > Tables, when there are no classifying factors.
c) There is a tidymodels alternative to summary called skim that will probably be useful as an alternative to the original summary, rather than a replacement?
d) We are going to be examining different ways of adding a group_by to R-Instat. I suggest there will be a variety of ways that we want to add group_by. I suggest we consider adding it here as an option for skim. That's because skim should produce tidy output (I assume). So including skim by a factor - i.e. thinking of it as an extension of one variable could be worth considering. This could be here, or as an option for 2 variables, see below.
Now the big one, namely 2 variables.
Currently, in 2 variables, (graph and summary) we are looking at one situation only. This is where we examine multiple variables of the same type, by a single variable. I would like to add further options - presumably through our usual radio buttons.
a) One reason that the ggpairs (that we examine in the correlations dialogue) does cope with factors as well as numeric. We are missing a great low-hanging fruit ignoring it!
b) So I suggest another option (button?) is perhaps called Pairs under 2 variable graphs, which just has a multiple receiver. And the data can at least be both numeric and factor (and we examine what happens with date, logical, character just to be sure. There is a wide range of options and this should be an exciting addition.
c) On the Summarise there are (at least) 2 interesting options. The first is for numeric and would be to put correlations (of pairs) as an option under 2 variables - maybe as well as multivariate. For simplicity (and because there are graphs as well as numerical summaries) I wonder about keeping it as a separate dialogue. There is another possibility here, namely to add (or change) to using the corrr package.
d) Then should we investigate the skim (tidymodels) with a by groups as an additional option for the 2 variables summarise. Then the multiple receiver can have data of mixed types and the second has to be (or is treated as) a group by variable? (If so, then one option for 3 variables will be skim with 2 factors as by variables.)
Beta Was this translation helpful? Give feedback.
All reactions