[ddj] we are beta launching our new tool, Xtraktr
ericnbarrett at gmail.com
ericnbarrett at gmail.com
Thu Sep 24 11:48:36 UTC 2015
Hi everyone,
I would like to share something we've been working on at JumpStart - Xtraktr
<http://xtraktr.jumpstart.ge/en>. Xtraktr is a data tool which makes
working with survey data easier, but not just survey data.
We had a local need in Georgia (the country) to make it easier for
organizations to publish their data since our work is so dependent on
access to datasets that describe the issues so important in our country. We
thought to reduce any obstacles preventing orgs from publishing their
survey and other data online. We also wanted to start to create a catalogue
of local datasets so we and anyone else could find them and use them.
Fortunately, ePaństwo Foundation <http://epf.org.pl/en/> invited us to a
hackathon last year where we got to start working on it. Then UNICEF in
Georgia <http://unicef.ge/> asked us to build a data portal using their
survey data, which allowed us to build out our tool as we (and they, of
course) wanted, but get funding for it :).
Xtraktr makes it easy to upload and explore a dataset, but really excels at
survey data and you can easily upload an SPSS or Stata file, as well as a
CSV or XLSX file. Once uploaded you can do a whole bunch of configuration
so tailor what users will see. We really couldn't find free tools that were
easy to use that made it easy to do something with an SPSS file. We wanted
journalists and NGOs to be able to do more with survey data than what they
currently do without huge fees.
The code for the application is open source
<https://github.com/JumpStartGeorgia/Xtraktr>, but you can use our hosted
version <http://xtraktr.jumpstart.ge/en> if you like. You can easily create
an account, and you can create an organizational account, too, once you
create your own account and invite other people to become members of the
organizations.
When you upload a dataset, you can choose which datasets are public or
private. Once 'extracted', Xtraktr creates a codebook of your questions and
answers (variables). You can edit the questions and answers, choose which,
if any, you don't want to include in the publicly explorable dataset, or
which answers you want to enable the users to exclude from the analysis.
Any dataset you upload automacally gets added to a restful API
<http://xtraktr.jumpstart.ge/en/api> with which you can interact with the
data using URLs and get JSON responses.
You can make a variable mappable by mapping it against a shape set. You can
upload new shape sets in GeoJSON format if you need to. This enables a
choropleth map of single variables.
You can weight questions by adding one or more weights and match weights to
one or more of your questions.
You can explore categorical variables as single variables (bar, pie, map,
table) or cross tabulate them (stacked bar, table). You can filter the
data.
All visualizations are widgets that you can download or embed in other
online media.
In addition, you can translate a dataset (questions and answers,
explanatory notes, categories and even the application interface) into as
many languages as you like. Xtraktr makes it easy be enabling a mass
changes spreadsheet that you then re-upload with the new translations and
any other edits you like to the texts.
Currently, you can only visualize categorical data - this is a result of
the initial survey focus. However, we are presently working to enable
visualizing and exploring relationships of integers/floats and that should
be ready in a month or two (hopefuly)!
To conclude this long explanation, I'd love to hear your thoughts and see
if this meets any of your case uses and if there are features you like or
things that you think we need to improve. Also, if you would like to
collaborate on this, we would also love to know that too :).
Looking forward to your feedback,
Eric
---
Eric Barrett
Executive Director
JumpStart Georgia <http://jumpstart.ge/en>
Check out Feradi Info <http://feradi.info/en>!
Check out Xtraktr <http://xtraktr.jumpstart.ge/en>!
Check out the Lari Explorer <http://lari.jumpstart.ge/en>!
Check out StoryBuilder <http://storybuilder.jumpstart.ge/>!
eric at jumpstart.ge
+995 595 13 92 69
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/data-driven-journalism/attachments/20150924/5be93347/attachment-0002.html>
More information about the data-driven-journalism
mailing list