Step 3: Data Enrichment¶

This section is not about feature creation (for an ML algorithm), but to enrich the asteroid dataframe with more, additional information.

In [1]:

# Import standard libraries
import os
import pathlib

# Import installed libraries
import pandas as pd

In [2]:

# Let's mount the Google Drive, where we store files and models (if applicable, otherwise work
# locally)
try:
    from google.colab import drive
    drive.mount('/gdrive')
    core_path = "/gdrive/MyDrive/Colab/asteroid_taxonomy/"
except ModuleNotFoundError:
    core_path = ""

In [3]:

# Read the level 1 dataframe
asteroids_df = pd.read_pickle(os.path.join(core_path, "data/lvl1/", "asteroids_merged.pkl"))

Bus classification to Main group¶

A great summary of asteroid classification schemas, the science behind it and some historical context can be found here. One flow chart shows the link between miscellaneous classification schemas. On the right side the flow chart merges into a general "main group". These groups are:

C: Carbonaceous asteroids
S: Silicaceous (stony) asteroids
X: Metallic asteroids
Other: Miscellaneous types of rare origin / composition; or even unknown composition like T-Asteroids

In [4]:

# Create a dictionary that maps the Bus Classification with the main group
bus_to_main_dict = {
                    'A': 'Other',
                    'B': 'C',
                    'C': 'C',
                    'Cb': 'C',
                    'Cg': 'C',
                    'Cgh': 'C',
                    'Ch': 'C',
                    'D': 'Other',
                    'K': 'Other',
                    'L': 'Other',
                    'Ld': 'Other',
                    'O': 'Other',
                    'R': 'Other',
                    'S': 'S',
                    'Sa': 'S',
                    'Sk': 'S',
                    'Sl': 'S',
                    'Sq': 'S',
                    'Sr': 'S',
                    'T': 'Other',
                    'V': 'Other',
                    'X': 'X',
                    'Xc': 'X',
                    'Xe': 'X',
                    'Xk': 'X'
                   }

In [5]:

# Create a new "main group class"
asteroids_df.loc[:, "Main_Group"] = asteroids_df["Bus_Class"].apply(lambda x:
                                                                    bus_to_main_dict.get(x, "None"))

In [6]:

# Remove the file path and Designation Number
asteroids_df.drop(columns=["DesNr", "FilePath"], inplace=True)

In [7]:

# Show the final data set for anyone who is interested ...
asteroids_df

Out[7]:

	Name	Bus_Class	SpectrumDF	Main_Group
0	1 Ceres	C	Wavelength_in_microm Reflectance_norm550n...	C
1	2 Pallas	B	Wavelength_in_microm Reflectance_norm550n...	C
2	3 Juno	Sk	Wavelength_in_microm Reflectance_norm550n...	S
3	4 Vesta	V	Wavelength_in_microm Reflectance_norm550n...	Other
4	5 Astraea	S	Wavelength_in_microm Reflectance_norm550n...	S
...	...	...	...	...
1334	1996 UK	Sq	Wavelength_in_microm Reflectance_norm550n...	S
1335	1996 VC	S	Wavelength_in_microm Reflectance_norm550n...	S
1336	1997 CZ5	S	Wavelength_in_microm Reflectance_norm550n...	S
1337	1997 RD1	Sq	Wavelength_in_microm Reflectance_norm550n...	S
1338	1998 WS	Sr	Wavelength_in_microm Reflectance_norm550n...	S

1339 rows × 4 columns

In [8]:

# ... and also the spectrum of Ceres
asteroids_df.loc[asteroids_df["Name"] == "1 Ceres"]["SpectrumDF"][0]

Out[8]:

	Wavelength_in_microm	Reflectance_norm550nm
0	0.44	0.9281
1	0.45	0.9388
2	0.46	0.9488
3	0.47	0.9572
4	0.48	0.9643
5	0.49	0.9716
6	0.50	0.9788
7	0.51	0.9859
8	0.52	0.9923
9	0.53	0.9955
10	0.54	0.9969
11	0.55	1.0000
12	0.56	1.0040
13	0.57	1.0056
14	0.58	1.0037
15	0.59	1.0036
16	0.60	1.0044
17	0.61	1.0071
18	0.62	1.0107
19	0.63	1.0113
20	0.64	1.0117
21	0.65	1.0127
22	0.66	1.0128
23	0.67	1.0124
24	0.68	1.0151
25	0.69	1.0160
26	0.70	1.0146
27	0.71	1.0178
28	0.72	1.0222
29	0.73	1.0216
30	0.74	1.0191
31	0.75	1.0179
32	0.76	1.0167
33	0.77	1.0149
34	0.78	1.0161
35	0.79	1.0176
36	0.80	1.0178
37	0.81	1.0196
38	0.82	1.0200
39	0.83	1.0164
40	0.84	1.0135
41	0.85	1.0140
42	0.86	1.0147
43	0.87	1.0151
44	0.88	1.0142
45	0.89	1.0146
46	0.90	1.0165
47	0.91	1.0181
48	0.92	1.0200

In [9]:

# Create Level 2 directory and save the dataframe
pathlib.Path(os.path.join(core_path, "data/lvl2")).mkdir(parents=True, exist_ok=True)

# Save the dataframe as a pickle file
asteroids_df.to_pickle(os.path.join(core_path, "data/lvl2/", "asteroids.pkl"), protocol=4)