[openspending-dev] OpenSpending Next visualization experiment from CSV files

[Daniel Fowler]

Hi all,

I've started some work with Rufus on visualizing spending data on the
web straight from CSV files instead of calling out to the OpenSpending
API.  It reuses a lot of the current visualization code for
OpenSpending and expects CSV files that are already grouped by some
number of dimensions.  For example, the CSV snippet below aggregates
UK Treasury spending 2004-2010 by COFOG Level 1 code and then by
Department code.

cofog_level1_code,dept_code,value
01,DFT004,425530000.0
03,NIO081,7368105000.0
03,NIE099,589154000.0

The first level of the visualization rendered using the above should
look something like this:

https://discuss.okfn.org/uploads/default/original/1X/b1ac20c5669c865d7bee52fe19e748f1d7ec6f40.png

I'm looking for ways to move forward with this project.  For instance,
given that the script that generates the aggregated CSVs could easily
generate them in denormalized form (e.g. actually showing department
names instead of just department codes), should there be any work on
the frontend to handle this same task for fully normalized tables?  Or
should I just render what I've been given?  I've posted a slightly
reworded version of this question here:

https://discuss.okfn.org/t/openspending-next-visualization-experiment/886

Comments welcome!

Cheers,

Dan