[okfn-za] Ideas for an initial group project - input wanted!

Mon Feb 11 12:13:56 UTC 2013

Hi Louisa,

Great suggestions. Comments inline below. BTW you accidentally did not
include the list in your response (need to hit reply-all - probably
should set the list to reply to list by default ...)

Also with this new influx of ideas I've booted a google doc that may
be easier to manage than a mailing list thread:

https://docs.google.com/a/okfn.org/document/d/1xi0uKiC0z0rpUQ2YVisLv0I61faYwy0eQtHS9uS7z4c/edit#

People can add there and then ping the list or just continue
discussion on thread.

On 9 February 2013 20:45, Louisa Potter <potter.louisa at gmail.com> wrote:
> Not sure what you mean by initial group project.
> To get newbies involved?
> To start new year?

More of the former but both are good. The other point is that I think
it useful to have a project for us to focus our efforts on a bit to
get the ball rolling.

> If I followed okfn more closely, I could probably make better suggestions.

The one's below are already excellent :-)

> Ideas I would like to see happen (and that could very well have zero
> to do with the type of project you are looking for):
>
> 1) pulling stats together for minors in prison: what was bail, how
> long has he/she been in there, how many times has court been postponed
> due to "lost" docket or absent key person

This sounds great. Who do you think would be interested in this
specifically? Are there journalists, or NGOs or others would be
especially interested?

Any idea where we could find the data? Is it all paper or is it
electronic? Does the government make it available. Could we focus on a
specific locality to start with?

> 2) using pharmacy sales to map minor "epidemics" in real time
> hay fever meds and/or flu meds sales moving from one town/province to next
> antidepressant sales in Gauteng (kidding mostly)
> lice shampoo sales in Hout Bay (serious)
> If pharms reluctant to provide info, results could be used by
> pharmacies to predict runs on specific products.(they must do this
> already??)

Another nice idea but I think we will have a lot of trouble getting
data (this will all be held by the pharmacy chains and getting it from
them will be hard (?))

> 3) degrees of separation mind-mappish type thingy for 500 most
> influential South Africans.
> Type in two names, and thingy finds the shortest routes between the
> two people (boards of directors, relatives, education, political
> events, tenders).
> Not "quite simple" as you requested but would be fascinating.

So we need the SA companies register including directors. theyrule for
South Africa [1]. Is the Companies Register available in machine
readable format (Aside: it would be great to have more info on South
African data in the Open Data Census: http://census.okfn.org/census/

[1]: http://www.theyrule.net/

> 4) Shifts in mining ownership in southern Africa. Including asking
> price for mines and price paid (sometimes much higher). Also include
> rarer critical metals and total % ownership - who is gaining monopoly
> on unusual metals.

This would be *very* interesting and have a nice connection with
http://OpenSpending.org/

Again, any ideas where we can get mining ownership info. Can you FoI
it or get it online?

> 5) something to illustrate potential coverage difference between
> current cell tower system and tower-less mesh cell-to-cell network
> system.

Not sure as exciting (or is this a big deal). I can imagine Steve Song
could be interested here.

> 6) national map of police station location juxtap with population
> I have in mind the Hout Bay police station for 20,000 people vs the
> eastern Baviaanskloof Nature Reserve police station for 20 people, 1/4
> of whom work in the station.

Again, very nice. Where

> I have more.

I've booted a gdoc: I suggest adding there and then mailing the list
https://docs.google.com/a/okfn.org/document/d/1xi0uKiC0z0rpUQ2YVisLv0I61faYwy0eQtHS9uS7z4c/edit#

> What was the question?

Finding a good first project for the OKFN-ZA group to work on. Some
suggested criteria:

* Interest
* Feasibility
  * Is there data
  * How technically complex is this to create
* Impact

Rufus

> Kind regards,
> Louisa