Make More Maps
I've been using #QGIS to visualise
Data Collection
... and I always collected the geo coordinates by hand. Looking up the place on Google maps, verified the aerial view, picked a main building, right click, click on the coordinates and paste into a csv file, separate the string "50.92828306087918, 6.929125426188075" into 50.928283 | 6.9291254
Now, I finally realised that ChatGPT can do the job in no time.
Let ChatGPT do the Job
Where are the Kenyan Research Institutes based? National Commission for Science and Technology (NACOSTI) has an official list of Kenyan #Research Institutes online - but there is no map showing us where they are based (and of course the list does not include coordinates).
For the following Kenyan research institutes I need the decimal geo coordinates of their physical address. Please create a csv file separated by semicolon:
Kenya Forestry Research Institute (KEFRI)
Kenya Marine and Fisheries Research Institute (KMFRI)
National Museums of Kenya (NMK)
British Institute in Eastern Africa
Kenya Medical Research Institute (KEMRI)
Kenya Industrial Research and Development Institute (KIRDI)
International Centre for Insect Physiolog y and Ecology (ICIPE)
Kenya Water Institute (KEWI)
The Nairobi Hospital
International Centre for Research in Agroforestry (ICRAF)
International Livestock Research Institute (ILRI)
Elizabeth Glaser Pediatric and AIDS Foundation
Institute of Primate Research
Institute of Energy Studies and Research
The Turkana Basin Institute
CAB International
Bioversity International
African Population and Health Research Centre
Mpala Research Centre
International Potato Centre (CIP)
Trendy Consulting International Limited
The old-fashioned way, twenty places would have taken me about half an hour. This is just a short example - imagine doing this with 430 German Universities ...
Ask for the Address first!
ChatGPT takes about twenty seconds to give you geo coordinates for the research institutes. Unfortunately, results for small institutions can be off by a few kilometers. In a testcase for all 37 universities in Berlin ChatGPT the mean distance between location on Google Maps and according to ChatGPT was 2,3 km. That is okay for a national map - but it is a no-go for a city map.
So, I asked ChatGPT for the address of the institution first and in a second step I asked for the geo coordinates for the address - which brought down the mean difference to just 600m. That is something I can live with.
Create short Labels
Research at the "International Centre for Insect Physiology and Ecology" is great but on the average map the name is consuming a lot of space - so I prefer ICIPE for short.
ChatGPT can do the work for us:
What are common abbreviations used for those institutions like ICIPE and KIMRI?
Within a few minutes we come up with a reasonable list of geo-referenced data:
I have a reasonable draft of an empty map of #Kenya:
Kenya is a very centralised state, and every single institute except the Mpala Research Centre is in or around Nairobi, so it is helpful to zoom in:
While this is far from being a perfect map, it is a starting point, and the technique really broadens up the horizon for GIS and data visualisation.
#GIS #maps #QGIS #ChatGPT