Where to next for SOFA

354 views
Skip to first unread message

Grant Paton-Simpson

unread,
Aug 3, 2013, 9:33:29 PM8/3/13
to sofasta...@googlegroups.com

The SOFA Statistics project could go in a number of different directions. Ideally, it would:

  • Add more chart types and more flexibility for graphical customisation
    (without compromising the SOFA goals of beautiful output and ease-of-use)
  • Add a comprehensive array of the most important statistical tests
    (without compromising the ease-of-use and learn-as-you-go goals)
  • Make it much easier to automate reporting
  • Make publishing reports to the web, and office formats, seamless and simple.

I only have limited time to develop SOFA at the moment, so I have to choose the top priorities. Here is what I think I should do:

  • Charting
    • Make it easy to export data so it’s ready for charting using spreadsheet charting tools
    • Provide brief documentation on how to use advanced tools like Matplotlib
  • Statistical Tests
    • Make it really easy to export data from SOFA ready for analysis in R
  • Report automation
    • Provide documentation so people can automate SOFA themselves using Python
  • Publishing
    • Add a plug-in for exporting to a document format

I should also solve the remaining bugs preventing Mac users from being able to export output as images.

What do people think about this direction? Drop me a line at gr...@sofastatistics.com.

Grant Paton-Simpson

unread,
Aug 3, 2013, 9:38:04 PM8/3/13
to sofasta...@googlegroups.com
Or post a comment here of course :-)

Alan Jones

unread,
Aug 5, 2013, 2:17:30 PM8/5/13
to sofasta...@googlegroups.com
I like the ideas of more chart types and flexibility and
statistical tests.  Not sure at what level i would go for the automated reporting at some point why customize to much vs just go to R.

I don't see the advantages to Exporting to some other document formats  For me the key is getting chars/graphs in a good quality image format that can be blown up if needed.

Exporting tables to excel or something,  and other basic JPG/PDF exports.

Exporting to Word or Open Document format may sound nice, but moving the images/tables in Word/LibreOffice may be easier better than exporting to that forma.

Grant Paton-Simpson

unread,
Aug 11, 2013, 2:29:25 AM8/11/13
to sofasta...@googlegroups.com
Hi Alan,

Thanks for your feedback. Sorry for the delay - I was waiting to see if I got much feedback from this forum or other sources so I could report back. Not a lot to report sadly :-(

I think I'll make fixing the Mac export images bug my next priority. And then think about things further at that point.

All the best,
Grant



On 06/08/13 06:17, Alan Jones wrote:
I like the ideas of more chart types and flexibility and
statistical tests.� Not sure at what level i would go for the automated reporting at some point why customize to much vs just go to R.

I don't see the advantages to Exporting to some other document formats� For me the key is getting chars/graphs in a good quality image format that can be blown up if needed.

Exporting tables to excel or something,� and other basic JPG/PDF exports.


Exporting to Word or Open Document format may sound nice, but moving the images/tables in Word/LibreOffice may be easier better than exporting to that forma.



On Saturday, August 3, 2013 8:33:29 PM UTC-5, Grant Paton-Simpson wrote:

The SOFA Statistics project could go in a number of different directions. Ideally, it would

  • Add more chart types and more flexibility for graphical customisation
    (without compromising the SOFA goals of beautiful output and ease-of-use)
  • Add a comprehensive array of the most important statistical tests
    (without compromising the ease-of-use and learn-as-you-go goals)
  • Make it much easier to automate reporting
  • Make publishing reports to the web, and office formats, seamless and simple.

I only have limited time to develop SOFA at the moment, so I have to choose the top priorities. Here is what I think I should do:

  • Charting
    • Make it easy to export data so it�s ready for charting using spreadsheet charting tools
    • Provide brief documentation on how to use advanced tools like Matplotlib
  • Statistical Tests
    • Make it really easy to export data from SOFA ready for analysis in R
  • Report automation
    • Provide documentation so people can automate SOFA themselves using Python
  • Publishing
    • Add a plug-in for exporting to a document format

I should also solve the remaining bugs preventing Mac users from being able to export output as images.

What do people think about this direction? Drop me a line at gr...@sofastatistics.com.

--
�
---
You received this message because you are subscribed to the Google Groups "sofastatistics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sofastatistic...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
�
�

Victor van Reijswoud

unread,
Aug 11, 2013, 4:27:54 AM8/11/13
to sofasta...@googlegroups.com
Dear Grant,

Requesting feedback in the holiday period in Europe, Africa and the
USA will not help you get a lot of responses..... You better repeat
this in October.

I like your improvement ideas, but I also like to bring up another
issue and that is to actively create a community around SOFA that will
take development tasks. I think the project needs more active people
on board (who program, create documentation, do marketing etc) so that
it is becomes more than a one-man-show. One several occasions I and
others have offered help, but this needs a plan and coordination.

Another thing is translations. In the French and Spanish domain the
interest for FOSS tools is high and we may want to tap into this.

Also an easy to use interface with the other statistical packages
needs to be created: e.g., SPSS, EpiInfo because SOFA is limited and
when people can not move ahead to these other packages easily when
they need more advanced analyses, they may not even start using it.

Finally, we need to determine the strengths and the weaknesses of SOFA
better in order to determine the target groups in more detail.

It is a bit of a different angle, but I hope this adds.

All the best,

Victor


On Sun, Aug 11, 2013 at 7:29 AM, Grant Paton-Simpson <gr...@p-s.co.nz> wrote:
> Hi Alan,
>
> Thanks for your feedback. Sorry for the delay - I was waiting to see if I
> got much feedback from this forum or other sources so I could report back.
> Not a lot to report sadly :-(
>
> I think I'll make fixing the Mac export images bug my next priority. And
> then think about things further at that point.
>
> All the best,
> Grant
>
>
>
> On 06/08/13 06:17, Alan Jones wrote:
>
> I like the ideas of more chart types and flexibility and
> statistical tests. Not sure at what level i would go for the automated
> reporting at some point why customize to much vs just go to R.
>
> I don't see the advantages to Exporting to some other document formats For
> me the key is getting chars/graphs in a good quality image format that can
> be blown up if needed.
>
> Exporting tables to excel or something, and other basic JPG/PDF exports.
>
> Exporting to Word or Open Document format may sound nice, but moving the
> images/tables in Word/LibreOffice may be easier better than exporting to
> that forma.
>
>
>
> On Saturday, August 3, 2013 8:33:29 PM UTC-5, Grant Paton-Simpson wrote:
>>
>> The SOFA Statistics project could go in a number of different directions.
>> Ideally, it would
>>
>> Add more chart types and more flexibility for graphical customisation
>> (without compromising the SOFA goals of beautiful output and ease-of-use)
>> Add a comprehensive array of the most important statistical tests
>> (without compromising the ease-of-use and learn-as-you-go goals)
>> Make it much easier to automate reporting
>> Make publishing reports to the web, and office formats, seamless and
>> simple.
>>
>> I only have limited time to develop SOFA at the moment, so I have to
>> choose the top priorities. Here is what I think I should do:
>>
>> Charting
>>
>> Make it easy to export data so it’s ready for charting using spreadsheet
>> charting tools
>> Provide brief documentation on how to use advanced tools like Matplotlib
>>
>> Statistical Tests
>>
>> Make it really easy to export data from SOFA ready for analysis in R
>>
>> Report automation
>>
>> Provide documentation so people can automate SOFA themselves using Python
>>
>> Publishing
>>
>> Add a plug-in for exporting to a document format
>>
>> I should also solve the remaining bugs preventing Mac users from being
>> able to export output as images.
>>
>> What do people think about this direction? Drop me a line at
>> gr...@sofastatistics.com.
>
> --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "sofastatistics" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sofastatistic...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
>
> --

Grant Paton-Simpson

unread,
Aug 11, 2013, 7:10:59 AM8/11/13
to sofasta...@googlegroups.com
Thanks for the reply Victor. More below:

On 11/08/13 20:27, Victor van Reijswoud wrote:
> Dear Grant,
>
> Requesting feedback in the holiday period in Europe, Africa and the
> USA will not help you get a lot of responses..... You better repeat
> this in October.
Good point.
> I like your improvement ideas, but I also like to bring up another
> issue and that is to actively create a community around SOFA that will
> take development tasks. I think the project needs more active people
> on board (who program, create documentation, do marketing etc) so that
> it is becomes more than a one-man-show. One several occasions I and
> others have offered help, but this needs a plan and coordination.
Any thoughts on how this could happen? I have limited resources to
devote to this at the moment but I am open to ideas. Of course, we will
probably have to wait till October but it is good to start thinking of
options now.
> Another thing is translations. In the French and Spanish domain the
> interest for FOSS tools is high and we may want to tap into this.
>
> Also an easy to use interface with the other statistical packages
> needs to be created: e.g., SPSS, EpiInfo because SOFA is limited and
> when people can not move ahead to these other packages easily when
> they need more advanced analyses, they may not even start using it.
I totally agree. It is possibly the most important factor of all in some
ways.

At the moment the Export Data option is only to export data to an xls
format but it could be extended quite easily to other formats. Probably
with a standard API like I made to allow database connections.
> Finally, we need to determine the strengths and the weaknesses of SOFA
> better in order to determine the target groups in more detail.
Do you have any particular thoughts about this?

Once again, thanks for providing feedback.

All the best,
Grant
>>> Make it easy to export data so it�s ready for charting using spreadsheet

Iain Gallagher

unread,
Mar 6, 2014, 9:04:51 AM3/6/14
to sofasta...@googlegroups.com

Hi Grant

I realise that this is well after the original post but... anyway.

Here are a couple of suggestions for SOFA.

Being able to chart values per category as points rather than e.g. a bar would be great. This would be similar ggplot2's geom_point() or geom_jitter() in R when used with a categorical x-axis variable as in the example plot (which hopefully comes up).

Secondly a post-hoc test after ANOVA (Tukeys HSD possibly) would be great.

Thanks.

Iain

Grant Paton-Simpson

unread,
Mar 8, 2014, 5:37:12 PM3/8/14
to sofasta...@googlegroups.com
Hi Iain,


On 07/03/14 03:04, Iain Gallagher wrote:

Hi Grant

I realise that this is well after the original post but... anyway.

Thanks for providing feedback.

Here are a couple of suggestions for SOFA.

Being able to chart values per category as points rather than e.g. a bar would be great. This would be similar ggplot2's geom_point() or geom_jitter() in R when used with a categorical x-axis variable as in the example plot (which hopefully comes up).
I've taken a note in my features list but I suspect it is not something which should be in an application as focused as SOFA (see http://www.sofastatistics.com/blog/using-sofa-alongside-other-statistics-packages/). But I am working on making it much easier to export from SOFA (along with value labels etc) into other programs that people can use to do more specialist analyses and charting.


Secondly a post-hoc test after ANOVA (Tukeys HSD possibly) would be great.
That has been asked for enough times and I can really see a need for it too. I'll vote it up in my features list. A friend of mine has thought quite a bit about how to make this for the project but he seems to be a bit snowed under. Adding this feature may well happen at some point and the goal would be to make it as helpful to users as possible (learn-as-you go being a top SOFA goal).

All the best,
Grant

Thanks.

Iain

Iain Gallagher

unread,
Mar 23, 2014, 10:38:22 AM3/23/14
to sofasta...@googlegroups.com
Hi Grant

Just a quick clarification on the first point. I don't mean that SOFA has to talk to R in anyway but rather that the ability to create a quick exploratory plot (beeswarm I believe it's called) could be added.

E.g.

# preliminaries
import numpy as np
from matplotlib import pyplot as plt
from scipy import stats

# some data
female = [63.8, 56.4, 55.2, 58.5, 64.0, 51.6, 54.6, 71]
male = [75.5, 83.9, 75.7, 72.5, 56.2, 73.4, 67.7, 87.9]

# the means
mean_female = np.mean(female)
mean_male = np.mean(male)
means = (mean_male, mean_female)

# create the plot
plt.figure()
plt.subplot(111)

ind = np.arange(len(means)) # create the x-axis
width = 0.5

plt.plot(ind, [male, female], color = 'grey', marker = 'o', ls='')
plt.ylabel('Kilograms', fontsize = 18)
plt.xticks(ind, ('Female', 'Male'), fontsize = 18)
plt.title('Weight by Sex', fontsize = 24)
plt.yticks(np.arange(0,110,20), fontsize = 14)

padding = [-0.5, 1.5, 0, 110] # [xmin, xmax, ymin, ymax]
plt.axis(padding)
plt.yticks(np.arange(0,110,10))

padding = [-0.5, 1.5, 0, 110] # [xmin. xmax, ymin, ymax]
plt.axis(padding)

# add a line for the means
plt.hlines(means, ind-0.2, ind+0.2, color='red')

plt.show() # oh, hello!


Hope it's not too presumptuous to add the above code to the post. The reason I think these are more useful than bar charts is that they let you see the structure of your data more easily. That might perhaps be in keeping with the more pedagogical aspects of SOFA i.e. getting people to see that the barchart can hide a lot of the detail of the data.

Cheers & thanks for SOFA.

Best

Iain

Grant Paton-Simpson

unread,
Mar 23, 2014, 2:40:45 PM3/23/14
to sofasta...@googlegroups.com
Nice - and there's nothing presumptuous about code offerings - I'll have a look later.
--

---
You received this message because you are subscribed to the Google Groups "sofastatistics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sofastatistic...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Grant Paton-Simpson

unread,
Mar 24, 2014, 2:35:12 PM3/24/14
to sofasta...@googlegroups.com
Hi Iain,

Thanks for that - I got it running but it seemed to need the same numbers in the two categories as written? Anyway, it looks interesting as an alternative to box plots for smaller datasets. No plans to add new chart types to SOFA at present but an interesting option that I understand better now with your example.

All the best,
Grant

On 24/03/14 03:38, Iain Gallagher wrote:

Chris Gale

unread,
Mar 24, 2014, 3:04:20 PM3/24/14
to sofasta...@googlegroups.com
Grant
Box plot visualixation and inter quartlie ranges are good ways of summarizing non parametric data. Which is most survey data.

Chris
--

---
You received this message because you are subscribed to the Google Groups "sofastatistics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sofastatistic...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Albert Law

unread,
Nov 10, 2015, 5:11:33 PM11/10/15
to sofastatistics
Hi Grant,

Greetings from Calgary, Canada. I am a new user of SOFA. Still learning how to import data from Excel file.
Great software. Thanks.

Albert

Grant Paton-Simpson

unread,
Nov 11, 2015, 2:56:53 AM11/11/15
to sofasta...@googlegroups.com
Thanks - glad you like it.
--
Reply all
Reply to author
Forward
0 new messages