[archived] Managing contact data in the Mapping of the Mappings

Tags: #<Tag:0x00007f426a101bd0>

(Jon Richter) #1

Continuing the discussion from Architektur von Geodateninfrastrukturen:

As @josefkreitmayer explicitly mentionned again that we have personal data publicly online and collected so tightly in the spreadsheet, I am really concerned about the privacy of those people.

I know @Giuliana gave passed the task to gather this info, but nevertheless we should respect the legal implications. Especially since we have no formal legal body available and the person who entered is totally responsible for what has been published there.

I propose to close the spreadsheet down and switch to “invite only / request” view from now on.

@alabaeye @toka @species ?

Providing privacy in public collected data
(Josef Kreitmayer) #2

I agree with the concern of @almereyda, but would suggest another measure
to resolve.

  • make a local copy of the online-document as it is and store it in our
  • make an online copy of the online-document as it is, and make it invite
  • once that is done, delete the columns with personal contact details in
    the public spreadsheet

here is a trello card for processing that issue:

other request: who is @toka?

(Jon Richter) #3

That’s perfectly enough.

He sings with us the song of data liberty, continuous delivery and full disclosure. An enigmatic apologist of freedom. As in speech, not in beer.

(Jon Richter) #6

It turns out the link on the blog still leads to the very same spreadsheet which contains the possibly sensitive data. I thought it had been copied and then cleared.
It turns out the revision history still offers all the data, also when we manually clear it.

Now people who use their bookmarks or our website can still access the compilation of addresses and phone numbers. Because it’s not that these numbers aren’t public, but their centralized compilation makes them vulnerable.

Unfortunately this also doesn’t work. There is a question of our same concern, but deleting and recreating doesn’t work either, as it changes the link.

But apparently it’s possible to Copy everything to the clipboard, restoring to the first version (or an appropriate one before all the details came), and pasting the cleaned version again. We should make a copy (with a new link) beforehand, sure.

Could you @josefkreitmayer take care of this, so I can mirror a copy on http://dat.transformap.co (which currently remains down) ? Thanks in advance.

(Josef Kreitmayer) #7

@almereyda Hello Jon, going back in history did not work, as the history-go-back is stored with the full history. So here is a link to the copy of a newly created spreadsheet without the contact details. https://docs.google.com/spreadsheets/d/1o3ghnT2jEsxlAGeG96gf1LrOgAVyhku5_e-lVzCrl4Y/edit?usp=sharing Please paste it, where the former link was.

If for internal use, the database with contact details is available. Here is the link to apply for database access. Acess is just granted if there is a clearly defined internally agreed purpose: https://docs.google.com/a/getactive.co.at/spreadsheets/d/10XVVr0hf05TiBmWqFIurjKrgLHTLcnq0frlH8Hw2t94/edit?usp=sharing

thx @toka for raising awareness. The issue seemed to be solved before, but due to some enabled filter, some adresses were still visible, and thx to the deeper research of @almereyda we also found the possible history leak.

(Jon Richter) #8

Yo @josefkreitmayer I’ve checked it. Looks quite good so far.

But please don’t delete the old spreadsheet for those who bookmarked or linked it elsewhere, because cool URIs don’t change, and just delete everything inside (which creates a new revision on top of the restored first one), and include a link to the new Mapping the Mappings of Alternative Economy without contact details with a small statement that the content has permanently moved to a new location. You may include the status code 301, too.

Muchas gracias in advance.

(Thomas Kalka) #9

I propose to

  • a) undelete the spreadsheet (Done)
  • b) change it permissions to be read and edited only by invited people (Done)
  • c) add a spreadsheet layer which displays only public data (Done)
  • d) publish this to the web (Done)
  • e) create a redirect an transformap.co to this data (transformap.co/data-sources.html ??)
  • f) change all links in posts to this redirect url

@almereyda, @josefkreitmayer what do you think ?

(Josef Kreitmayer) #10

@almereyda @toka

I restored the document, and put it non public.
@toka is invited to edit now.

please @almereyda & @toka do as you think best. I do not dedicate more time to that.
Including I am fine with all you want to do with it.

thank you!

(Thomas Kalka) #11

@josefkreitmayer Thank you.

a public version is now available at

a csv at

the public version is on a separate worksheet in the same spreadsheet, copying the public data via


etc references, then published to web

(Adrien Labaeye) #12

Thanks for taking care of this! But I think the whole point was to let anyone enter a new initiative or? I was about today to share the link to the French commoners so they can complete it…

(Thomas Kalka) #13

Adding more alternatives will still be possible using a form …

(Adrien Labaeye) #16

I would suggest that we still display some more columns like a) existing mapping system, googlemaps/OSM, description. It would be useful for people to navigate the inventory.
Who is preparing the form?

(Thomas Kalka) #17

Yes, this needs more work …
I added more collumns
Json output should be possible also, see http://pipetree.com/qmacro/blog/2013/10/sheetasjson-google-spreadsheet-data-as-json/

(Adrien Labaeye) #18

there is currently some bug on the public version… weird: https://docs.google.com/spreadsheets/d/1gyqgHDoEj-w8LEdcPBs_J1FMtOf0GHFzZcHPtAQMorI/pubhtml?gid=1879121710&single=true

Would we want to move towards a wiki table rather? like this: http://hackerspaces.org/wiki/List_of_Hacker_Spaces

(Thomas Kalka) #19

This is a bug in the css / formatting.

If you just copy the data into some spreadsheet, you get it right.

Basically using a semantic wiki is a good idea.
I would prefer to keep a spreadsheet, until someone has a bigger time slot available to investigate usage of wikidata / wikibase or semantic mediawiki.


after opening one hidden column it seems to look right now

next small steps to improve user experience could be

  • enable json output
  • hack some facetted view / table sorter based on the json data to present it

(Adrien Labaeye) #20

@toka: how does the uuid code get generated? do I have to do anything after adding a map?

(Jon Richter) #21

I have added a small formula which you just need to extend further down. I’ve gotten the idea to look for it while thinking of the geocoder the OKFN Timemap Sheets use.

@toka Could you update the code so it generates small letters?

Update 06.03.2014: This issue isn’t resolved, yet.

The old link, which has been publicly accessible online the latest since beginning of May last year, does still show the contact details. It should be swapped with the current clean table by

  • reverting to the first available version,
  • deleting all additional tables (Is anyone really using them?) and
  • publishing it to the web.

I tried the lattest with the current clean one. After reinitializing dat.transformap.co with an empty dat repository (dat init on the server) I could then easily run the following:

FILE=`date -Iminutes`'_mmm_clean.csv'
if [ ! -d $DAT ]; then mkdir -p $DAT; fi && cd $DAT
dat clone $PROTO$SOURCE && cd $SOURCE
wget -O $FILE https://docs.google.com/spreadsheets/d/1o3ghnT2jEsxlAGeG96gf1LrOgAVyhku5_e-lVzCrl4Y/export?format=csv && cat $FILE | dat import --csv
dat push $PROTO$SOURCE

Unfortunately this still doesn’t consider changes and only appends everything ever and ever again. We may want to use daff for the intermediary step. (dat and daff :wink: )
The wget part could also be exchanged with curl, if one wants to omit storing a csv locally next to the dat. If such a clone exists already, one changed into the $SOURCE folder and issued dat pull $PROTO$SOURCE instead for updating the local copy.

Further on, all improvements regarding UUIDs, thanks @toka for the menu entry, however that came into existence, have only been applied on the old link. We still have no workflow to securely share the UUIDs between both instances.
Again, using Google Docs as a database, as appealing as it may have been and continues to be, turns out to be a bad choice. This will have to be one main objective for an initial 2015 monkeys call.

(Thomas Kalka) #22

@alabaye it is a script in googles app script

@jon will put the code in some gist so anybody can change it

I have started to play with semantic mediawiki, which seems to be the best tool, to collect this data collectively.

(Jon Richter) #23

Our experience tells us that the evaluation of “the best tool” can be quite tensing, once one starts to consider multiple options. As long as we only look at one solution, of course it will be the best.

The semantic MediaWiki does proposedly fit in for Ontology Design, but then it also concurrents with Ontowiki.
If the discussion is about managing many maps, the ideal interface should be a map itself. This is then another question to answer.

(Thomas Kalka) #24

hm. words.

“best tool” = “currently best option to get the work done, without much programming”

But yes, it feels well to have a rationale behind a decision.