Setting MLP for regression of 5 dimensional data

I have a set of data that consists of 280k datapoints and each datapoint has 5 parameters (latitude, longitude, unixtime, wind speed and wind bearing). I use regression to fit a model to this dataset so I can output the wind speed and wind bearing at locations between locations in the original dataset.

I have done this to a degree, but upon testing I have found the outputs are pretty far off from expected.
For example: From the dataset at coords: -37.7276, 144.9066, 1483174800, (lat, long, unix time). I should get windspeed = 16.6 and bearing = 193. But instead when I put the coords into the trained model I get [189.1, 310.24]. I assume windspeed = 189 and bearing = 310 because this is how i setup my y-data for input.

Any help would be greatly appreciated