I was never able to keep a diary going longer than 6 weeks, so it's no surprise that I haven't blogged for ages.  But don't worry, 'walk 'n talk-ers', work has been progressing.  Not as fast as I would have liked (Phil and James, too, no doubt), but progress has been made.  Things did take a little while to pick up after Christmas, but eventually some really helpful people got to know about our project and have been encouraging friends and relations to join in.

So, the tally stands at:

Walking interviews - 7 done; 1 arranged.

Sedentary interview - 3 done.

Doubles - 5 completed; 1 to finish (but this does include one walking interview where the PDA failed).

Counting on my fingers tells me that is 17 people interviewed, plus I have 3 more people to contact after Easter.  (I don't actually have 17 fingers - I had to take my socks off.)

In other news, I have been receiving a steady stream of interview transcipts from Jon, who is doing a good job of this essential, but slightly tedious task!  I am now trying to find time to listen to the recording while checking the transcript for accuracy, trying to remember what was said in the bits we can't hear very well and assessing the noise level.  I then go through them a second time to 'smoothe' them slightly - i.e. take out some of the 'ums' and 'errs'.  This is potentially controversial as we are trying to reflect what people say about the area.  On the other hand, things that make sense verbally look very strange written down, and without hearing the inflection in the voice, meaning may be changed.  Also, it seems that few of us (and I was horrified to discover that I am very much included in this, given that talking is one of the key requirements of my job!) can speak in proper sentences!!  While the interviews are always fine at the time, all the 'ums' and repetitions while people think of the right word makes following the thread of the conversation difficult - not to mention rather boring.  Transcripts are then sent to the interviewee to ensure that they are a correct reflection of what was said - and it also gives the respondent a chance to take out any comments that may have slipped out inadvertently.  It doesn't happen very often, but sometimes if we feel comfortable with  someone (and I hope you all feel happy being interviewed by me!) we forget that the recorder is running....  Anyway, once I get the OK from the interviewee, I select some of the stories to go on the website, and Phil does something technical which attaches the text to the GPS location et viola! - it appears on the website.  Look out after Easter for more....

Well, it keeps me out of mischief.

So, as Easter approaches, I wonder 'where did the time go?'  Well, actually, no.  I really think - 'yippee, holiday!'  Happy Easter, all - I'll be back in a couple of weeks.... xxx

Well, I was snowed under with marking for most of February (just finished the last of the first batch this morning - more coming in on Friday) and I was frustrated because something which almost worked wasn't quite working.

I've been trying to find a way of getting the transcripts up on the website in such a way that you could follow the conversation around as the walk was undertaken.  The first step was to attach the transcripts (which our transcriber Jon has broken into 10 second chunks) to the GPS points recorded during the walk (every 10 seconds).  I did this by putting each 10 seconds of text into a cell in Microsoft Excel, along with a number from 0-3 which Jane has generated for each ten seconds to indicate the level of background noise on the recording (buses going by being 3).  In ArcGIS I joined the Excel tables to the GPS logs and then exported these to KML using a script called Export to KML which is produced by the City of Portland Bureau of Planning.  This allowed me to use the text from the transcript as the KML Feature Label - as well as use the background noise number to refer to one of a series of little picture files I created to represent increased volume.

All this I'd done before and the logs look fine in Google Earth or on the Google Maps site, both of which automatically generate a table of contents based on the Feature Label for each point.  I wanted, however, to have this table of contents embedded on the Rescue Geography website.  This apparently simple thing has taken me absolutely ages to figure out and I have to acknowledge Mike Williams, who has an excellent tutorial on using Google Maps code.  He has also written a very, very clever little piece of script called EGeoXML which allows you to do all kinds of things to take data from your KML files and present it on your website in a variety of ways.

The upshot of all this (yes, I know, it's tedious, but it's a field diary of the developing method) is that for Blair Kesseler's interview (which Jane did back in October) we now have the full transcript with a numbered table which allows you to click and have balloons open on the map in the order of the text - following the conversation.  We also have a series of edited highlights on his main page, with a drop down box (again generated using EGeoXML) allowing you to jump to the stories Blair told about different locations.

While I'm here I should probably say something about CP's question to James mentioned in his last post.  KML is now a fairly standard format for exchanging geographic data on the web.  Anyone who wants to can take our KML files and do other things with them - you're only limited by your expertise in understanding the (basic) programming language underneath them, rather than whether you have access to GIS software like ArcGIS.  While there are other ways of getting this lovely data onto the web, Google do make it very easy for the amateur dabbler in code such as myself.  I will confess that I am uncomfortable about using Google for all of this stuff for a number of reasons:
i) they are an evil multinational
ii) the free map service could be withdrawn at any time
iii) questions of futureproofing - the site currently uses the latest version of KML and Google Maps, but will this code be supported by whatever the successor applications are in, say, ten years time.

There are, unfortunately, practicalities.  We don't have an ArcIMS server to run all this stuff directly from my GIS to the web.  Even if we did, the license to run Ordnance Survey data (i.e. to indicate where the streets are) on ArcIMS is utterly extortionate.  It's also not very shareable.  So KML is definitely the way forward.  But why use Google and not one of user-generated map sites?

Essentially while I'm pretty good at this kind of stuff, I'm incredibly dependent on people like Mike Williams posting stuff on the web explaining how to write/use the code and, for the moment at least, the vast majority of people who know about these things are writing about KML as it relates to Google.  I may one day figure out how to live a Google-free existence, but for the moment it's a bit like having a car - I cycle as much as possible and someday I may get rid of the car, but it's awfully handy having one when you want to go to the shops...