i did it myself, after a very long and difficult manipulation of european nuts1 shapefiles:) if anyone needs this map i can share the stata code and the mapfiles

I see you have done it yourself, but I was going to recommend getting the European file from eurostat (ec.europa.eu/eurostat) and then extracting Turkey only.  I use MapWindow software (free from mapwindow.org) which makes this easy - you can just filter the shapes to get those matching NUTS_ID Like "TR%" and STAT_LEVEL_ = 1 and then export those shapes to a new shapefile.

Note I tried out this method to check if worked, so if you want further information (or the shapefile I exported for comparison with yours) just let me know.

http://www.mapwindow.org/

http://ec.europa.eu/eurostat/web/gisco/geodata/reference-data/administrative-units-statistical-units/nuts#nuts13

thank you. I tried a much difficult method, converted the whole dbf file to csv and then deleted all x y coordinates other than turkey nuts level regions, then conerted this csv file to dbf again... Your method is more simple

I downloaded MapWindow and compared both maps and they are same

actually the program is doing the same thing that I did by hand, since I deleted all the other coordinates other than I need

I would recommend anyone to be careful when using NUTS1 level or NUTS 2 or 3 for spatial econometrics. The presence of the Marmara Sea requires that you hand -edit your neighborhood matrix. Beware of transport connections between the surrounding regions.

Also, since there are only very few number of NUTS 1 or 2 regions, I find it not very useful to use spatial econometrics, as you will quickly run into multicollinearity and will come out with very high R squares. Rather, one could create a variable which represents these spatial relationships, and then run a normal regression. As Luke Anselin and James La Sage says, you do not always have to use a complex spatial econometrics model, but do the right thing. Otherwise I am not an expert on it. In addition, the spatial error factor applies mostly to continuities of cities, to my opinion on Adana-Mersin and Istanbul Kocaeli, and maybe combining these two regions would be a rather better option than introducing more variables.

If you need ARCGIS shapefile, send me an email and maybe I can find one for you. Best regards.

Hi,

you can get NUTS 1 level shapefile by using ARCview program. If you have province (NUTS 3) level shapefile, you can merge provinces and get NUTS 1 level shapefile.

Best...

Hi,

If you need help for spatial panel data analysis (mekansal panel veri analizi) and also if you need help to get Turkey's NUTS1 and NUTS2 shape files (Türkiye'nin NUTS1 ve NUTS2 shape file dosyaları), you may contact with me.

I can advise you on your academic studies, especially on panel data econometrics.

Pls visit my web site www.tunakanduran.com

Best regards,

Dr. Tunakan DURAN

www.tunakanduran.com