Saving sets of cluster coordinates

Question

I had a set of coordinates and plotted it as scatter graph. Then DBSCAN was applied to the points so the points that are closed together form a cluster (Note: the black dots are being classified as noise).

Now I wanted to find the centroid of the cluster (by adding the x & y coordinates together and average it out). I am able to obtain the set of coordinates (that is grouped into its respective cluster) and it looks like this:

[[10. 75.]
 [11. 74.]
 [11. 75.]
 [12. 73.]
 [12. 74.]
 [12. 75.]]
[[34. 49.]
 [34. 50.]
 [35. 48.]
 [35. 49.]
 [35. 50.]
 [36. 48.]
 [36. 49.]
 [36. 50.]]
[[43. 78.]
 [43. 79.]
 [43. 80.]
 [43. 81.]
 [44. 78.]
 [44. 79.]
 [44. 80.]
 [44. 81.]
 [45. 78.]
 [45. 79.]
 [45. 80.]
 [45. 81.]
 [46. 78.]
 [46. 79.]
 [46. 80.]
 [46. 81.]
 [47. 78.]
 [47. 79.]
 [47. 80.]
 [47. 81.]
 [48. 79.]
 [48. 80.]]
[[53. 63.]
 [53. 64.]
 [53. 65.]
 [54. 63.]
 [54. 64.]
 [54. 65.]
 [54. 66.]
 [55. 63.]
 [55. 64.]
 [55. 65.]
 [55. 66.]
 [56. 63.]
 [56. 64.]
 [56. 65.]
 [56. 66.]]
[[ 72. 115.]
 [ 73. 114.]
 [ 73. 115.]
 [ 73. 116.]
 [ 73. 117.]
 [ 73. 118.]
 [ 74. 113.]
 [ 74. 114.]
 [ 74. 115.]
 [ 74. 116.]
 [ 75. 113.]
 [ 75. 114.]
 [ 75. 115.]
 [ 75. 116.]
 [ 75. 117.]
 [ 76. 113.]
 [ 76. 114.]
 [ 76. 115.]
 [ 76. 116.]
 [ 76. 117.]
 [ 77. 115.]
 [ 77. 116.]]
[[79. 56.]
 [79. 57.]
 [79. 58.]
 [79. 59.]
 [80. 57.]
 [80. 58.]
 [80. 59.]]

...where each [[ & ]] marks the start and end of a cluster. So I am wondering if there is a way that I can save the set of cluster coordinates in a way that each group of cluster coordinates are saved seperately into its individual txt file (to make it easier to perform the averaging calculation later)?

If these are numpy arrays (as opposed to python lists) you might look at [`np.save`/`np.load`](https://docs.scipy.org/doc/numpy/reference/generated/numpy.save.html) (see: https://stackoverflow.com/questions/51837627/why-does-pickle-take-so-much-longer-than-np-save/51845434#51845434) — wbadart, Apr 17 '19 at 14:59

VnC · Answer 1 · 2019-04-17T14:45:05.637

If you have the following 2 clusters:

clusters_2d = [
    [
        [10, 75,],
        [11, 74,],
        [11, 75,],
        [12, 73,],
        [12, 74,],
        [12, 75,],
    ],
    [
        [34, 49,],
        [34, 50,],
        [35, 48,],
        [35, 49,],
        [35, 50,],
        [36, 48,],
        [36, 49,],
        [36, 50,],
    ]
]

In order to save them as separate files you can iterate over the coordinates:

Using pickle

import pickle

for i, x in enumerate(clusters_2d):
    pickle.dump( x, open( f"cluster_{i}.pkl", "wb" ) )

Then in order to read one of those files:

cluster1 = pickle.load( open( "cluster_0.pkl", "rb" ) )

Using txt

for i, x in enumerate(2d_clusters):
    with open(f'cluster_{i}.txt', 'w') as f:
        for item in x:
            f.write(f"{item}\n")

You can either save them in a pickle file or txt or whatever you like.

Saving sets of cluster coordinates

1 Answers1