What I've learned in 2 years of using the Thai schools spreadsheet

Due to weather and arthritis, I’ve had lots of time over the last two years to do map stuff. So far, I’ve gone through 13 (of 76) provinces and would like to share some things I’ve learned. Most provinces have 200 to 400 schools and usually take about one to two months to do. Most of that time is adding roads and rivers, and redrawing rivers and roads that were not done so well in the past due to poor imaging. I don’t really care much about the schools, but the database is a gold mine of village names. Adding a wat and school (polygons), and a village name to the map improves an otherwise featureless group of residential roads.

The usefulness of the data varies among the provinces. Some use the village name for all the schools and wats, some usually use the wat name (usually also the village name), a few have no geographical information, and most are a mix. So far, I’ve added names for at least one of the three features (village, school, wat) for 7000 entries. The total named features added is about twice that. BTW, there were 950 entries with incorrect GPS data.

In the past, my mapping was random or just the areas I visited on a road trip. Following the data base organizes the process by covering most of the province while going through the school names. As Russ mentioned, there are lots of missing roads. They are impossible to find without some framework like the database. The current province I’m working on is Yasothon, and about half of the roads and villages were not drawn.

Thai school names are usually the village name plus a few syllables that may be archaic and untranslatable, but usually just ‘science’ or ‘progress’. If nothing else, my Thai reading is way better and I seldom need Thai2English.

I usually cross check names with Geonames, but never take data from it. When I find a village that is already named, it is often from Geonames and I change the source to ‘https//data.go.th’.

This is of course all pretty tedious, but most is just cut and paste. I also have found several shortcuts. Most of the time is spent fixing or adding roads and rivers, not entering data. (FB doesn’t do any fixing or add water).

If anyone is interested in trying this, please let me know. I have enough experience with it to be able to help.

Tom

Thanks for the write-up and the long hours you’ve spent adding data, Tom. I wasn’t aware of the trove of data at https//data.go.th but now that I am, I will refer to it when I come to a town that has no posted names along the highway, my usual source for them. I have often used temple names to determine the names of towns when that name includes the word “ban”. Same for schools.

Cheers from Alaska,

Dave

Thanks Tom for the effort to bring names to the map.

Suppose that the missing roads are not just a matter of time for having the gaps filled up by Facebook team and other mappers.

Saying that: I’m going to meet part of the Facebook team in Italy this weekend during the State of the Map conference. If you have any topics to discuss with them, please let me know.

Regarding the effort to create polygons: I personally would not create the polygons for temples/schools. Besides it being quite a huge effort, often you can’t be certain from aerial what is exactly belonging to them. A simple node is much quicker and gives the same details to the map.

Stephan

Hi Stephan,

Of course, these roads would be drawn eventually. But it’s years in and there are still lots to go.

What I would like to say to FB is please try to draw the roads more carefully, and to take a moment to fix poorly drawn roads they encounter. And to try to find a GPS trace nearby to align the image. Please say ‘hi’ to "DrishT. I envy you going to Italy! I would eat too much.

Polygons of schools and wats are no more difficult to draw than polygons of anything else. Also, my ‘method’ requires polygons, because the buffer can hold one node (village) and one polygon (school). Saves a lot of work.

Here is an image of Thailand in dots representing schools. The brilliant script is by Mishari.

https://photos.google.com/photo/AF1QipOgRCygBntuuhE4Vzrtkf_cCM5OJkstZfrgFg9N