This entry also appears here. 

Note: names mentioned other than my own are pseudonyms.

It’s been an eventful 2 weeks. The Saturday after arriving I went on a big mountain bike ride looping west then south of PP. It was with expats, some with very nice expensive carbon bikes. It was 70km. I rented a bike. I crashed it. I was going about 40km/h on the dirt, hit a bump and slid down the (gravel) road. Ouchies. 10 days later I still have some healing to go on my right knee and elbow, but the grazes on my right hand have pretty much done sealing up. I was so scared when I got injured because I heard that infections can be hard to beat here, so I pre-emptively took a course of antibiotics that you’re supposed to take for infections, I think it was a good idea as I definitely got swelling on my hand. There’s no infection anymore but I’m still laying on the Betadine every day. 

 The first week at work was a bit of a trainwreck as well. I had to cope with the language barrier and the somewhat less centralised way of storing data that they use. They do not have a file server. There is no spatial warehouse. The source and date of most reference datasets are unknown. There are multiple copies of the same data, stored in shapefiles with each different map. There were no templates for map layouts (I made 6, portrait/landscape for A4, 3 and 0). No one was using the latest version of our mapping software (ArcGIS). They all have old computers with one small screen each. The main GIS guy actually has the smallest screen. The two GIS assistants (who do other community surveying and development work and both have degrees in economics) actually have widescreen perhaps 20 or 21” displays. As an example of issues with the data, I had to put our rivers dataset in Google Earth and digitise new banks because they were all out by 5-10m, and there is an extra island now in the Mekong River (unless the person making the original just left that island out). Possibly the rivers were digitised from a small scale data source, I don't know as there was no metadata. We also have no cadastre dataset, even though we’re doing a land rights project! Jim 1, my friend from my GIS Master’s program who lives in PP, said that we can get this if we apply to the relevant ministry, although I’ll have to do it as someone from EWB rather than the local NGO of course, as they’re suspended… Progress so far has been, as mentioned, to make 6 templates, for example this A4 landscape one. 
 

---

The background image in the marginalia comes from a professional photographer who took a few photos along the abandoned/being rehabilitated railway in PP. This was to do with their MIT (mapping, infrastructure and titling) project where they have been monitoring the impact of the railway development on those living in the railway corridor. The others were grateful for the templates although their use necessitated an upgrade to ArcGIS 10 as I used this to make them. This required Windows XP SP 3 being installed on one of the guy’s computers, which took about 2 days of stuffing around as automatic update wanted to install all this other stuff beforehand. There is still one of the guys to go, he was initially suspicious of the templates, thinking that so many unique maps are made that they wouldn’t be of use. He was more receptive when I said that the templates could be used for regular maps. I think someone just needs to stick their neck out and say this is how we can make our operation more efficient and professional, although I shouldn’t perhaps be too forthright given I’m new. 

Another piece of progress has been to initiate a spatial warehouse. I have the same set up as at Jim's Engineering where there is a base geodatabase with feature datasets based on the ISO topic classification.

 I have begun to put data in these categories, although Jim 2, the main GIS guy at STT, has not given me all of the data yet. There is a whole lot that has come from a guy called Jim 3, although it’s all in shapefiles (half of which have the Indian 1960 UTM Zone 48N projection for some reason, the one we use is WGS 1984 UTM Zone 48N). There is no metadata with these files. I think it reflects the enterprise of GIS here… we simply make do with whatever we can get our hands on, just having something is better than nothing at the moment so issues of quality/currency are not so important. I’m hoping to take things up a level in this respect, it will increase the credibility of our work. This will probably entail a meeting with ‘Jim 3’ to find out where his data is from. Then we can put it in the all-important Sources section of our maps. The spatial warehouse also has BaseData and ProjectData folders for things such as DEMs and imagery, not that we have much of this though. At the moment we just georeference an image from Google Earth (GE) whenever we need to do some house digitising etc. The cool thing about upgrading to Arc10 though is that there is a Bing Maps aerial imagery baselayer WMS that can be used. Hogs bandwidth though. At least it removes the bother of having to georeference GE imagery all the time. 

Another method of creating some order is to create a common job folder structure. I made a Generic Job Folder (see pictures).

 Extra subfolders will probably be added as things progress, but it’s good to have a start. I’m so glad I had some familiarity with these methods of organisation from my work at Jim's Engineering! Amidst all of this administration work I have managed to make 5 maps. These are improvements of maps already made by Jim 2, which Jane 1 wanted to be, in her words, ‘sexed up’. Basically they show eviction sites in PP, sites excluded from the systematic land titling project and locations of communities threatened with eviction in PP. It was fun to make these maps. It would be nice to show you but I will have to discuss what I can share about such documents with the higher ups before releasing anything here. Initially we were using a roads dataset either from Jim 3 or someone else. This contained roads going right through the lake in the north of the Khan of Sensok. So we decided to use data from Openstreetmap, which is more extensive anyway. For example there were no roads on the island in Chamkarmon in the old dataset. 

Three other forms of upskilling I have told them about are dynamic text, data driven pages (DDPs) and creating tables in excel/word that can simply be pasted as an image into the layout view of ArcMap. You can see such a table at the bottom of the last map. This had been made ever so painstakingly using ruler guides for each line and column by Jim 2 (a practice also employed by the GIS assistants Jim 4 and Jim 5). Individual text elements were used for each of the values in their tables. I wonder if this may have been an artefact of their having to work with Khmae language in the past and this not working/displaying in Excel so it was easiest just to insert individual text elements for each table value and arrange them all manually in the layout view. Anyway, I showed them how to put the table together in Excel then put it in Word for the finishing touches (such as getting rid of the grid lines), before finally pasting into ArcMap as an image… which can be resized to suit.

I also showed them about dynamic text, which is implemented in much more of a major way in Arc10. I showed Jim 2 particularly all of the dynamic text elements in the templates, such as to automatically update the scale, date, coordinate system, file name etc etc. We are also using the Map Document window to contain the title, subtitle, authors and source information. This is then fed to the relevant areas of the template by dynamic text. Lastly I showed Jim 2 briefly how to use data driven pages to automatically centre the map on the 8 Khans of PP, for example when we’re making individual maps of them for the eviction sites. He said he used bookmarks for this, so he was pleased with the automatic printing function of DDPs where you can let it churn through all records in a particular layer automatically, and give you individual PDFs, for example. I think with these sorts of improvements we can make our work more efficient and spend more time on quality and analysis. 

Speaking of spatial analysis, Jane 1, the head honcho at STT, said she wants to analyse the proximity of areas excluded from systematic land titling to both eviction sites and those threatened with eviction. The first step of this is contained in the above maps, but some proper spatial analysis is needed, as far as I understand what she is thinking, to determine how well we can use the areas excluded to predict evictions. There is more discussion needed about this though as I’m not fully clear on what she wants me to do. 

It’s the water festival/independence day celebration this week so we only worked Monday and Tuesday. Yesterday we had an annual planning meeting and I presented about SPSS, the other main form of contribution I’ll be making to STT. This is a statistics program used to test things like the statistical significance of a difference in scores between two groups on a particular variable. I imagine we’ll do a whole lot of correlation analyses although there has been no discussion about research questions yet so I’m not sure. I introduced the program with a T-Test example but it appeared not many understood. It appeared that few had done much with statistics apart from your regular descriptive statistics in Excel. So I will probably have to run a series of presentations on basic statistics before we can start using SPSS. I was half wondering this, so brought my first year introduction to statistics workbook from Flinders… looks like I’ll be putting it to use. 

In other news I went out with an English teacher all day today. Nice way to participate in the Independence Day festivities! It all started on the dancefloor at Riverhouse last Friday. Just like it should. Still haven’t bought a motorcycle, but have moved on from the idea of a Daelim and, possibly therefore, my efforts to say “I am in solidarity with the attainable aspirations of the poor majority” (you MUST read the blog Stuff Expat Aid Workers Like). I am going to get a dirt bike, then I can do big rides. Like a Honda XR 250 or Baja. Although I saw an ATV fanging along Sihanouk last week. Basically anything that goes goes in PP. I’m loving all the old Jeeps too. I have, however, purchased a bicycle. Anywhere else it would be considered a girl’s bike (ok, probably just Australia/outside Asia) as it has no horizontal top tube. But it’s very useful with a rack on the back, basket front, dyno light, mudguards and a very relaxed riding position
 
Other photos from around the place are below.