I am fairly new to machine learning, but what I am trying to do is predict energy output of a wind turbine based on a data set of wind speed and direction, and ideally turbulence and shear.

Do you know what type of data I need? I know I need a database of wind speed, direction, and possibly turbulence and shear. My questions are as follows:

Thank you!!

  • Please given some more information: Do you want to have measurement or modelling data? Do you need it for a specific location or for a wider area? Does it need to be US data or can it be also European data? Dec 1, 2016 at 8:58
  • 1
    To question 1: When you are interested in modelling data, [coastDat2](www.coastDat.de) is a source for European modeling data. To question 2: Yes, you can load them into MATLAB (you can load every text file into MATLAB). However, writing the file's content sensibly into variables needs a bit preparation. The format is well described here. To question 4: what did you find out so far on your own to answer this question? Dec 1, 2016 at 9:03
  • Do you have experience in importing such text files into MATLAB or generally importing/parsing such text files? Dec 1, 2016 at 9:05
  • 1
    It might be reasonable to have a look at Earth Science(earthscience.stackexchange.com). Dec 1, 2016 at 10:18
  • 1
    I marked this as closed because there are too many questions, and most of which are related to machine learning instead of open data. Additionally, there are already answered questions on this site regarding wind speed and power, see for example opendata.stackexchange.com/q/4412/1511
    – philshem
    Dec 3, 2016 at 8:24

1 Answer 1


You can download wind data in many different temporal resolutions from the Weather Underground, wunderground.com.

To get the data from Berlin-Tegel Airport TXL, I can do


and get

TimeCEST,TemperatureC,Dew PointC,Humidity,Sea Level PressurehPa,VisibilityKm,Wind Direction,Wind SpeedKm/h,Gust SpeedKm/h,Precipitationmm,Events,Conditions,WindDirDegrees,DateUTC
12:20 AM,9.0,1.0,58,1029,-9999.0,North,5.6,-,N/A,,Clear,360,2016-05-04 22:20:00
12:50 AM,9.0,1.0,58,1029,-9999.0,North,5.6,-,N/A,,Clear,360,2016-05-04 22:50:00
1:20 AM,8.0,2.0,66,1029,-9999.0,North,5.6,-,N/A,,Clear,350,2016-05-04 23:20:00
1:50 AM,5.0,1.0,76,1029,-9999.0,Variable,3.7,-,N/A,,Clear,0,2016-05-04 23:50:00
2:20 AM,6.0,1.0,70,1029,-9999.0,Variable,3.7,-,N/A,,Clear,0,2016-05-05 00:20:00
2:50 AM,6.0,1.0,70,1029,-9999.0,Variable,1.9,-,N/A,,Clear,0,2016-05-05 00:50:00
3:20 AM,6.0,2.0,76,1029,-9999.0,Variable,3.7,-,N/A,,Clear,0,2016-05-05 01:20:00

I know this data source from the R package weatherData. Haven't used this data much. Cannot tell how accurate this data is, qhat the quotas/rate limits are, if there are gaps, etc.

Not the answer you're looking for? Browse other questions tagged or ask your own question.