Numerical vs Categorical data
37 views
Skip to first unread message
Stathis Kanterakis
Sep 13, 2012, 11:57:05 AM9/13/12
to isaf...@googlegroups.com, Timo Wittenberger
Hello,
Is there a way to define whether a column has numerical or categorical data?
Do these correspond to "Factor Value" and "Characteristic" respectively?
Is there a way to guarantee that a column is numerical?
Please excuse me if this is a trivial question or is stated elsewhere.
Kind regards,
Stathis
Susanna-Assunta Sansone
Sep 13, 2012, 12:00:24 PM9/13/12
to isaf...@googlegroups.com, Stathis Kanterakis, Timo Wittenberger
Hi Stathis,
I guess you refer to the ISAcreator and want to set the allowed values;
if this is the case then the answer is yes, you can set the allowed
values via the ISAconfigurator when creating your configuration.
Susanna
--Susanna-Assunta Sansone, PhD
Principal Investigator
Team Leader
Projects: www.isacommons.org|www.mibbi.org|www.biosharing.org
University of Oxford Tel: +44(0)1865 610622
Oxford e-Research Centre Fax: +44(0)1865 610624
7 Keble Road, Oxford skype: susanna-a.sansone
OX1 3QG, UK uk.linkedin.com/in/sasansone
--
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> To unsubscribe from this group, send email to
> isaforum+u...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/isaforum?hl=en-GB
>
> Visit the ISAtools website at http://isa-tools.org
I guess you refer to the ISAcreator and want to set the allowed values;
if this is the case then the answer is yes, you can set the allowed
values via the ISAconfigurator when creating your configuration.
Susanna
--Susanna-Assunta Sansone, PhD
Principal Investigator
Team Leader
Projects: www.isacommons.org|www.mibbi.org|www.biosharing.org
University of Oxford Tel: +44(0)1865 610622
Oxford e-Research Centre Fax: +44(0)1865 610624
7 Keble Road, Oxford skype: susanna-a.sansone
OX1 3QG, UK uk.linkedin.com/in/sasansone
--
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> To unsubscribe from this group, send email to
> isaforum+u...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/isaforum?hl=en-GB
>
> Visit the ISAtools website at http://isa-tools.org
Eamonn Maguire
Sep 13, 2012, 12:29:23 PM9/13/12
to isaf...@googlegroups.com, Stathis Kanterakis, Timo Wittenberger
When you select integer as the type, a regex will be set for this field to accept only integers. Same for double but it will allow 0.0 as well as 0. However, it will only highlight the incorrect value in red on edit. A better way would be to keep the highlighting on always to show which fields are not valid. This can be added relatively easily though. If you want this, then add it to the issues on github as a feature request, and we will see when we can add it in.
--
Categorical data can be defined either as a normal string, a list (where you can define the values within that list) or as an ontology term.
Hope this answers your question.
Many thanks,
Eamonn
For more options, visit this group at
http://groups.google.com/group/isaforum?hl=en-GB
Visit the ISAtools website at http://isa-tools.org
--
--
You received this message because you are subscribed to the Google
Groups "ISAforum" group.
To post to this group, send email to isaf...@googlegroups.com
To unsubscribe from this group, send email to
For more options, visit this group at
http://groups.google.com/group/isaforum?hl=en-GB
Visit the ISAtools website at http://isa-tools.org
------------------------------------------------------------- Eamonn Maguire Lead Software Engineer - ISA tools
University of Oxford Tel: +44(0)1865 610788 Oxford e-Research Center 7 Keble Road, Oxford skype: maguire1986 OX1 3QG, UK -------------------------------------------------------------
Stathis Kanterakis
Sep 17, 2012, 4:35:51 AM9/17/12
to Eamonn Maguire, isaf...@googlegroups.com, Timo Wittenberger
Hi Eamonn and Susanna,
Thank you for your replies. This question was actually brought up when Timo asked if it is possible to determine from the column headers whether a column contains numerical or categorical data (for downstream statistical analysis). Perhaps you can get that from the configuration xml file. But there is no link from the investigation file (i_Investigation.txt) back to the configuration file (transcription_micro.xml). So how would one know where to look? Have you had a similar question before?
Thanks again.
Stathis
Philippe
Sep 17, 2012, 8:58:58 AM9/17/12
to isaf...@googlegroups.com, Stathis Kanterakis, Eamonn Maguire, Timo Wittenberger
Hi Stathis, Timmo,
It could be done indeed by looking up column headers in the
ISAconfiguration that was used to save the file.
ISAcreator now saves the configuration information used to create and
last save an ISA archive so you should be able to obtain all the
necessary information.
...
Investigation Public Release Date ""
Comment [Created with configuration] ""
Comment [Last Opened With Configuration] "isaconfig-default_v2011-02-18
copy"
...
As Eamonn replied, if the field datatype is of List type or Ontology
Entry type, you are dealing with categorical values while finding
integer or double indicates numerical data.
One thing to be aware is that users can add and define additional fields
beyond those declared in any given ISAconfiguration.
In this situation, it means ISAcreator has been used to create new
fields. Those can either be 'qualitative' or 'quantitative'. If the
latter, a Unit fields will be automatically added, immediately to the
right of the new field.
This feature can therefore be used as a test to cast columns as
'numerical' or 'categorical'.
Best wishes
Philippe
> <http://www.mibbi.org>|www.biosharing.org
> <http://www.biosharing.org>
> <tel:%2B44%280%291865%20610622>
> <http://uk.linkedin.com/in/sasansone>
> <mailto:isaforum%2Bunsu...@googlegroups.com>
> <mailto:isaforum%2Bunsu...@googlegroups.com>
>
> University of Oxford Tel: +44(0)1865 610788 <tel:%2B44%280%291865%20610788>
It could be done indeed by looking up column headers in the
ISAconfiguration that was used to save the file.
ISAcreator now saves the configuration information used to create and
last save an ISA archive so you should be able to obtain all the
necessary information.
...
Investigation Public Release Date ""
Comment [Created with configuration] ""
Comment [Last Opened With Configuration] "isaconfig-default_v2011-02-18
copy"
...
As Eamonn replied, if the field datatype is of List type or Ontology
Entry type, you are dealing with categorical values while finding
integer or double indicates numerical data.
One thing to be aware is that users can add and define additional fields
beyond those declared in any given ISAconfiguration.
In this situation, it means ISAcreator has been used to create new
fields. Those can either be 'qualitative' or 'quantitative'. If the
latter, a Unit fields will be automatically added, immediately to the
right of the new field.
This feature can therefore be used as a test to cast columns as
'numerical' or 'categorical'.
Best wishes
Philippe
> <sa.sa...@gmail.com <mailto:sa.sa...@gmail.com>> wrote:
>
> Hi Stathis,
> I guess you refer to the ISAcreator and want to set the allowed
> values; if this is the case then the answer is yes, you can set
> the allowed values via the ISAconfigurator when creating your
> configuration.
> Susanna
>
> --Susanna-Assunta Sansone, PhD
>
> Principal Investigator
> Team Leader
>
> Projects: www.isacommons.org
> <http://www.isacommons.org>|www.mibbi.__org
>
> Hi Stathis,
> I guess you refer to the ISAcreator and want to set the allowed
> values; if this is the case then the answer is yes, you can set
> the allowed values via the ISAconfigurator when creating your
> configuration.
> Susanna
>
> --Susanna-Assunta Sansone, PhD
>
> Principal Investigator
> Team Leader
>
> Projects: www.isacommons.org
> <http://www.mibbi.org>|www.biosharing.org
> <http://www.biosharing.org>
> <tel:%2B44%280%291865%20610622>
> Oxford e-Research Centre Fax: +44(0)1865 610624
> <tel:%2B44%280%291865%20610624>
> <http://uk.linkedin.com/in/sasansone>
> --
>
>
> On 13/09/2012 16:57, Stathis Kanterakis wrote:
>
> Hello,
> Is there a way to define whether a column has numerical or
> categorical data?
> Do these correspond to "Factor Value" and "Characteristic"
> respectively?
> Is there a way to guarantee that a column is numerical?
> Please excuse me if this is a trivial question or is stated
> elsewhere.
>
> Kind regards,
> Stathis
>
> --
> --
>
> You received this message because you are subscribed to the
> Google
> Groups "ISAforum" group.
> To post to this group, send email to
> isaf...@googlegroups.com <mailto:isaf...@googlegroups.com>
>
>
> On 13/09/2012 16:57, Stathis Kanterakis wrote:
>
> Hello,
> Is there a way to define whether a column has numerical or
> categorical data?
> Do these correspond to "Factor Value" and "Characteristic"
> respectively?
> Is there a way to guarantee that a column is numerical?
> Please excuse me if this is a trivial question or is stated
> elsewhere.
>
> Kind regards,
> Stathis
>
> --
> --
>
> You received this message because you are subscribed to the
> Groups "ISAforum" group.
> To post to this group, send email to
> To unsubscribe from this group, send email to
> isaforum+unsubscribe@__googlegroups.com
> <mailto:isaforum%2Bunsu...@googlegroups.com>
> For more options, visit this group at
> http://groups.google.com/__group/isaforum?hl=en-GB
> <http://groups.google.com/group/isaforum?hl=en-GB>
>
> Visit the ISAtools website at http://isa-tools.org
>
>
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> <mailto:isaf...@googlegroups.com>
>
> Visit the ISAtools website at http://isa-tools.org
>
>
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> To unsubscribe from this group, send email to
> isaforum+unsubscribe@__googlegroups.com
> <mailto:isaforum%2Bunsu...@googlegroups.com>
> For more options, visit this group at
> http://groups.google.com/__group/isaforum?hl=en-GB
> <http://groups.google.com/group/isaforum?hl=en-GB>
>
> Visit the ISAtools website at http://isa-tools.org
>
>
>
>
> --
>
>
> -------------------------------------------------------------
> Eamonn Maguire
> Lead Software Engineer - ISA tools <http://www.isa-tools.org>
>
> Visit the ISAtools website at http://isa-tools.org
>
>
>
>
> --
>
>
> -------------------------------------------------------------
> Eamonn Maguire
>
> University of Oxford Tel: +44(0)1865 610788 <tel:%2B44%280%291865%20610788>
> Oxford e-Research Center
> 7 Keble Road, Oxford skype: maguire1986
> OX1 3QG, UK
> -------------------------------------------------------------
>
>
>
> 7 Keble Road, Oxford skype: maguire1986
> OX1 3QG, UK
> -------------------------------------------------------------
>
>
>
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> To unsubscribe from this group, send email to
> isaforum+u...@googlegroups.com
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> To unsubscribe from this group, send email to
Stathis Kanterakis
Sep 17, 2012, 9:29:49 AM9/17/12
to isaf...@googlegroups.com, Eamonn Maguire, Timo Wittenberger
Hi Philippe,
You are right, the configuration is stored in the investigation file but as I mentioned before, there is no way to know which of the xml files in that configuration was used to define fields in a given ISA.
I guess the best way is then to check for a "Unit" column next to a factor, as you mentioned.
Kind regards,
Stathis
Philippe
Sep 17, 2012, 9:41:29 AM9/17/12
to isaf...@googlegroups.com, Stathis Kanterakis, Eamonn Maguire, Timo Wittenberger
Hi Stathis,
On 17/09/2012 14:29, Stathis Kanterakis wrote:
> Hi Philippe,
> You are right, the configuration is stored in the investigation file but
> as I mentioned before, there is no way to know which of the xml files in
> that configuration was used to define fields in a given ISA.
Really?
If you are looking at a study sample files (s_ .txt) then reading the
'StudySample.xml' file should give you the list of 'characteristics,
parameter values and factor values' and their attributes.
Then, when considering Assays, by forming a look-up on 'Measurement
Type' and 'Technology type' in the ISA investigation file and reading
the configuration files (e.g. looking for:
"<measurement term-label="histone modification profiling"
term-accession="" source-abbreviation="OBI"/>
<technology term-label="nucleotide sequencing" term-accession=""
source-abbreviation="OBI"/>" ,
you should be in a position to obtain the list of descriptors specified
for a given assay type.
Let me know if I have missed something.
all the best
Philippe
> <mailto:isaf...@googlegroups.com> <mailto:isaf...@googlegroups.com
> > <mailto:isaforum%2Bunsu...@googlegroups.com
> <mailto:isaforum%252Buns...@googlegroups.com>>
> > <mailto:isaforum%2Bunsu...@googlegroups.com
> <mailto:isaforum%252Buns...@googlegroups.com>>
On 17/09/2012 14:29, Stathis Kanterakis wrote:
> Hi Philippe,
> You are right, the configuration is stored in the investigation file but
> as I mentioned before, there is no way to know which of the xml files in
> that configuration was used to define fields in a given ISA.
If you are looking at a study sample files (s_ .txt) then reading the
'StudySample.xml' file should give you the list of 'characteristics,
parameter values and factor values' and their attributes.
Then, when considering Assays, by forming a look-up on 'Measurement
Type' and 'Technology type' in the ISA investigation file and reading
the configuration files (e.g. looking for:
"<measurement term-label="histone modification profiling"
term-accession="" source-abbreviation="OBI"/>
<technology term-label="nucleotide sequencing" term-accession=""
source-abbreviation="OBI"/>" ,
you should be in a position to obtain the list of descriptors specified
for a given assay type.
Let me know if I have missed something.
all the best
Philippe
> > <mailto:eamo...@gmail.com <mailto:eamo...@gmail.com>>> wrote:
> >
> > When you select integer as the type, a regex will be set for this
> > field to accept only integers. Same for double but it will
> allow 0.0
> > as well as 0. However, it will only highlight the incorrect
> value in
> > red on edit. A better way would be to keep the highlighting on
> > always to show which fields are not valid. This can be added
> > relatively easily though. If you want this, then add it to the
> > issues on github as a feature request, and we will see when we can
> > add it in.
> >
> > Categorical data can be defined either as a normal string, a list
> > (where you can define the values within that list) or as an
> ontology
> > term.
> >
> > Hope this answers your question.
> >
> > Many thanks,
> >
> > Eamonn
> >
> >
> > On 13 September 2012 17:00, Susanna-Assunta Sansone
> > <sa.sa...@gmail.com <mailto:sa.sa...@gmail.com>
> >
> > When you select integer as the type, a regex will be set for this
> > field to accept only integers. Same for double but it will
> allow 0.0
> > as well as 0. However, it will only highlight the incorrect
> value in
> > red on edit. A better way would be to keep the highlighting on
> > always to show which fields are not valid. This can be added
> > relatively easily though. If you want this, then add it to the
> > issues on github as a feature request, and we will see when we can
> > add it in.
> >
> > Categorical data can be defined either as a normal string, a list
> > (where you can define the values within that list) or as an
> ontology
> > term.
> >
> > Hope this answers your question.
> >
> > Many thanks,
> >
> > Eamonn
> >
> >
> > On 13 September 2012 17:00, Susanna-Assunta Sansone
> > <sa.sa...@gmail.com <mailto:sa.sa...@gmail.com>
> <mailto:isaf...@googlegroups.com>>
> > To unsubscribe from this group, send email to
> > isaforum+unsubscribe@__googlegroups.com
> <http://googlegroups.com>
> > To unsubscribe from this group, send email to
> > isaforum+unsubscribe@__googlegroups.com
> > <mailto:isaforum%2Bunsu...@googlegroups.com
> <mailto:isaforum%252Buns...@googlegroups.com>>
> > For more options, visit this group at
> > http://groups.google.com/__group/isaforum?hl=en-GB
> > <http://groups.google.com/group/isaforum?hl=en-GB>
> >
> > Visit the ISAtools website at http://isa-tools.org
> >
> >
> > --
> > --
> >
> > You received this message because you are subscribed to
> the Google
> > Groups "ISAforum" group.
> > To post to this group, send email to
> isaf...@googlegroups.com <mailto:isaf...@googlegroups.com>
> > <mailto:isaf...@googlegroups.com
> > http://groups.google.com/__group/isaforum?hl=en-GB
> > <http://groups.google.com/group/isaforum?hl=en-GB>
> >
> > Visit the ISAtools website at http://isa-tools.org
> >
> >
> > --
> > --
> >
> > You received this message because you are subscribed to
> the Google
> > Groups "ISAforum" group.
> > To post to this group, send email to
> isaf...@googlegroups.com <mailto:isaf...@googlegroups.com>
> <mailto:isaf...@googlegroups.com>>
> > To unsubscribe from this group, send email to
> > isaforum+unsubscribe@__googlegroups.com
> <http://googlegroups.com>
> > To unsubscribe from this group, send email to
> > isaforum+unsubscribe@__googlegroups.com
> > <mailto:isaforum%2Bunsu...@googlegroups.com
> <mailto:isaforum%252Buns...@googlegroups.com>>
> > For more options, visit this group at
> > http://groups.google.com/__group/isaforum?hl=en-GB
> > <http://groups.google.com/group/isaforum?hl=en-GB>
> >
> > Visit the ISAtools website at http://isa-tools.org
> >
> >
> >
> >
> > --
> >
> >
> > -------------------------------------------------------------
> > Eamonn Maguire
> > Lead Software Engineer - ISA tools <http://www.isa-tools.org>
> >
> > University of Oxford Tel: +44(0)1865 610788
> <tel:%2B44%280%291865%20610788> <tel:%2B44%280%291865%20610788>
> > http://groups.google.com/__group/isaforum?hl=en-GB
> > <http://groups.google.com/group/isaforum?hl=en-GB>
> >
> > Visit the ISAtools website at http://isa-tools.org
> >
> >
> >
> >
> > --
> >
> >
> > -------------------------------------------------------------
> > Eamonn Maguire
> > Lead Software Engineer - ISA tools <http://www.isa-tools.org>
> >
> > University of Oxford Tel: +44(0)1865 610788
> > Oxford e-Research Center
> > 7 Keble Road, Oxford skype: maguire1986
> > OX1 3QG, UK
> > -------------------------------------------------------------
> >
> >
> >
> > --
> > --
> >
> > You received this message because you are subscribed to the Google
> > Groups "ISAforum" group.
> > To post to this group, send email to isaf...@googlegroups.com
> > isaforum+u...@googlegroups.com
> > 7 Keble Road, Oxford skype: maguire1986
> > OX1 3QG, UK
> > -------------------------------------------------------------
> >
> >
> >
> > --
> > --
> >
> > You received this message because you are subscribed to the Google
> > Groups "ISAforum" group.
> > To post to this group, send email to isaf...@googlegroups.com
> > http://groups.google.com/group/isaforum?hl=en-GB
> >
> > Visit the ISAtools website at http://isa-tools.org
>
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> <mailto:isaf...@googlegroups.com>
> To unsubscribe from this group, send email to
> isaforum+u...@googlegroups.com
> >
> > Visit the ISAtools website at http://isa-tools.org
>
> --
> --
>
> You received this message because you are subscribed to the Google
> Groups "ISAforum" group.
> To post to this group, send email to isaf...@googlegroups.com
> <mailto:isaf...@googlegroups.com>
> To unsubscribe from this group, send email to
Stathis Kanterakis
Sep 17, 2012, 10:28:23 AM9/17/12
to isaf...@googlegroups.com, Eamonn Maguire, Timo Wittenberger
Yep. Understood. Thanks
Reply all
Reply to author
Forward
0 new messages