CODEWITHSUNDEEP
pythonarraysfunctionnumpymatrix
Marcus K.
Marcus K.

Reputation: 403

How do I create a 2D-array from calculation of a formula using two arrays as input?

I am trying to calculate how many hours of sunlight each latitude of Earth receives every year. I have a formula which calculates this, that uses two arrays as input, the day of the year, and the latitude. What I want to do is to use the formula to create a 2D-array with each day on one axis and each latitude on another, but I am unsure how to do it since my formula will only return the dot product of the two arrays rather than the cross product, making the result only a single axis. My code looks something like this:

Latitude = np.linspace(1,90,90)
Days = np.linspace(1,365,365)
P = np.arcsin(0.39795 * np.cos(0.2163108 + 2 * np.arctan(0.9671396 * np.tan(0.00860 * (Days - 186)))))

daylight = 24 - (24 / np.pi) * np.arccos((np.sin(0.8333 * np.pi / 180) + np.sin(Latitude * np.pi / 180) * np.sin(P)) / (np.cos(Latitude * np.pi / 180) * np.cos(P)))

What I what is then to calculate the formula "daylight" but to create a resulting 2D-array such that each value of P is multiplied with each value of Latitude, creating a 90x365 matrix.

Upvotes: 0

Views: 1126

Answers (3)

amzon-ex
amzon-ex

Reputation: 1744

If this is the core problem statement:

What I what is then to calculate the formula "daylight" but to create a resulting 2D-array such that each value of P is multiplied with each value of Latitude, creating a 90x365 matrix.

Then, just do the outer product of the two arrays, from your formula,

_P = np.sin(P)
_Lat = np.sin(Latitude * np.pi / 180)
PLat = np.outer(_P,  _Lat)

np.outer used in this form will multiply each element of the first array (size m) with each element of the second (size n), yielding a m x n array.
Read the docs.

After you've created PLat, the outer product matrix, add/multiply it with the other scalars.

Upvotes: 1

Joshua
Joshua

Reputation: 86

In Python you can create arrays like this:

a = [i for i in range(5)]

And you can also put multiple of these inside each other:

a = [[x*y for x in P] for y in Latitude]

And of course here you could insert your formula:

daylight = [[24 - (24 / np.pi) * np.arccos((np.sin(0.8333 * np.pi / 180) + np.sin(y * np.pi / 180) * np.sin(x)) / (np.cos(y * np.pi / 180) * np.cos(x))) for x in P] for y in Latitude]

Also this result contains a lot of nan values. From the context I would assume that these are polar nights/days. You could perhaps do some sort of np.arccos(min(1,max(-1,...)) to take care of these.

Upvotes: 1

quacodas
quacodas

Reputation: 273

Try numpy.meshgrid. This will create two 2D-arrays for latitude and days for you:

lat2d, days2d = np.meshgrid(Latitude, Days)

Those arrays will have shape (365, 90) and one axis represents one variable (so in lat2d, all rows will be the same and in days2d all columns will be the same). You can then use those arrays as in your formula.

Upvotes: 1

Related Questions