SpaceX Falcon 9 First Stage Landing Analysis¶
Exploring and Preparing Data¶
Estimated time: About 70 minutes (may vary depending on your pace)
In this notebook, I will explore and prepare the Falcon 9 launch data to better understand the factors that influence landing success. The ability to reuse the first stage is a key reason for SpaceX's cost advantage, and predicting landing outcomes is both a technical and business challenge.
I will perform exploratory data analysis (EDA) and feature engineering to prepare the data for future modeling.
What makes a Falcon 9 landing successful?¶
Below, I will also highlight some examples of unsuccessful landings, which are important for understanding the challenges SpaceX faces.
Note: Some unsuccessful landings are planned, such as controlled descents into the ocean.
My Objectives¶
- Perform exploratory data analysis (EDA)
- Engineer features for modeling
- Visualize key patterns and relationships in the data
Import Libraries and Define Helper Functions¶
I will import the following libraries for data analysis and visualization.
In [1]:
# import piplite
# await piplite.install(['numpy'])
# await piplite.install(['pandas'])
# await piplite.install(['seaborn'])
In [2]:
# pandas is a software library written for the Python programming language for data manipulation and analysis.
import pandas as pd
#NumPy is a library for the Python programming language, adding support for large, multi-dimensional arrays and matrices, along with a large collection of high-level mathematical functions to operate on these arrays
import numpy as np
# Matplotlib is a plotting library for python and pyplot gives us a MatLab like plotting framework. We will use this in our plotter function to plot data.
import matplotlib.pyplot as plt
#Seaborn is a Python data visualization library based on matplotlib. It provides a high-level interface for drawing attractive and informative statistical graphics
import seaborn as sns
Exploratory Data Analysis¶
First, let's read the SpaceX dataset into a Pandas dataframe and print its summary
In [3]:
!pip install js
Collecting js
Downloading js-1.0.tar.gz (2.5 kB)
Preparing metadata (setup.py) ... done
Collecting fanstatic (from js)
Downloading fanstatic-1.4-py3-none-any.whl (53 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 53.9/53.9 kB 700.9 kB/s eta 0:00:00
Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from js) (67.7.2)
Collecting WebOb>=1.2 (from fanstatic->js)
Downloading WebOb-1.8.7-py2.py3-none-any.whl (114 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 115.0/115.0 kB 2.8 MB/s eta 0:00:00
Building wheels for collected packages: js
Building wheel for js (setup.py) ... done
Created wheel for js: filename=js-1.0-py3-none-any.whl size=2883 sha256=cd2b649dc2218e254fe15af84846fd6b31083ecda0c7f7205e563e78e5da3f47
Stored in directory: /root/.cache/pip/wheels/77/dc/68/d580eaea0ef6137289a09dabe661f5df14951f5c29c08d443d
Successfully built js
Installing collected packages: WebOb, fanstatic, js
Successfully installed WebOb-1.8.7 fanstatic-1.4 js-1.0
In [4]:
from google.colab import drive
drive.mount('/content/drive')
Mounted at /content/drive
In [5]:
sns.set(rc={'figure.figsize':(14, 6)})
URL = "/content/drive/MyDrive/IBM Data Science Capstone/dataset_part_2.csv"
df=pd.read_csv(URL)
df.head(5)
Out[5]:
| FlightNumber | Date | BoosterVersion | PayloadMass | Orbit | LaunchSite | Outcome | Flights | GridFins | Reused | Legs | LandingPad | Block | ReusedCount | Serial | Longitude | Latitude | Class | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2010-06-04 | Falcon 9 | 6104.959412 | LEO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0003 | -80.577366 | 28.561857 | 0 |
| 1 | 2 | 2012-05-22 | Falcon 9 | 525.000000 | LEO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0005 | -80.577366 | 28.561857 | 0 |
| 2 | 3 | 2013-03-01 | Falcon 9 | 677.000000 | ISS | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0007 | -80.577366 | 28.561857 | 0 |
| 3 | 4 | 2013-09-29 | Falcon 9 | 500.000000 | PO | VAFB SLC 4E | False Ocean | 1 | False | False | False | NaN | 1.0 | 0 | B1003 | -120.610829 | 34.632093 | 0 |
| 4 | 5 | 2013-12-03 | Falcon 9 | 3170.000000 | GTO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B1004 | -80.577366 | 28.561857 | 0 |
First, let's try to see how the FlightNumber (indicating the continuous launch attempts.) and Payload variables would affect the launch outcome.
We can plot out the FlightNumber vs. PayloadMassand overlay the outcome of the launch. We see that as the flight number increases, the first stage is more likely to land successfully. The payload mass is also important; it seems the more massive the payload, the less likely the first stage will return.
In [6]:
sns.catplot(y="PayloadMass", x="FlightNumber", hue="Class", data=df, aspect = 5)
plt.xlabel("Flight Number",fontsize=20)
plt.ylabel("Pay load Mass (kg)",fontsize=20)
plt.show()
We see that different launch sites have different success rates. CCAFS LC-40, has a success rate of 60 %, while KSC LC-39A and VAFB SLC 4E has a success rate of 77%.
Next, let's drill down to each site visualize its detailed launch records.
TASK 1: Visualize the relationship between Flight Number and Launch Site¶
Use the function catplot to plot FlightNumber vs LaunchSite, set the parameter x parameter to FlightNumber,set the y to Launch Site and set the parameter hue to 'class'
In [7]:
# Use catplot to plot FlightNumber vs LaunchSite
sns.catplot(x='FlightNumber', y='LaunchSite', hue='Class', data=df,aspect=1)
# Set the title of the plot
plt.title('Flight Number vs Launch Site')
# Show the plot
plt.show()
In [8]:
# Plot a scatter point chart with x axis to be Flight Number and y axis to be the launch site, and hue to be the class value
# Plot scatter point chart
sns.scatterplot(x='FlightNumber', y='LaunchSite', hue='Class', data=df)
# Set the title of the plot
plt.title('Flight Number vs Launch Site')
# Show the plot
plt.show()
Now try to explain the patterns you found in the Flight Number vs. Launch Site scatter point plots.
TASK 2: Visualize the relationship between Payload and Launch Site¶
We also want to observe if there is any relationship between launch sites and their payload mass.
In [9]:
# Plot a scatter point chart with x axis to be Pay Load Mass (kg) and y axis to be the launch site, and hue to be the class value
# Set the style of the visualization
sns.set_style("whitegrid")
# Create a scatter plot
sns.scatterplot(x='PayloadMass', y='LaunchSite', hue='Class', data=df)
# Set the title of the plot
plt.title('Payload Mass vs Launch Site')
# Show the plot
plt.show()
In [10]:
# Display the DataFrame
print(df[['PayloadMass', 'LaunchSite', 'Class']].to_string(index=False))
PayloadMass LaunchSite Class 6104.959412 CCAFS SLC 40 0 525.000000 CCAFS SLC 40 0 677.000000 CCAFS SLC 40 0 500.000000 VAFB SLC 4E 0 3170.000000 CCAFS SLC 40 0 3325.000000 CCAFS SLC 40 0 2296.000000 CCAFS SLC 40 1 1316.000000 CCAFS SLC 40 1 4535.000000 CCAFS SLC 40 0 4428.000000 CCAFS SLC 40 0 2216.000000 CCAFS SLC 40 0 2395.000000 CCAFS SLC 40 0 570.000000 CCAFS SLC 40 1 1898.000000 CCAFS SLC 40 0 4707.000000 CCAFS SLC 40 0 2477.000000 CCAFS SLC 40 0 2034.000000 CCAFS SLC 40 1 553.000000 VAFB SLC 4E 0 5271.000000 CCAFS SLC 40 0 3136.000000 CCAFS SLC 40 1 4696.000000 CCAFS SLC 40 1 3100.000000 CCAFS SLC 40 1 2257.000000 CCAFS SLC 40 1 4600.000000 CCAFS SLC 40 1 5500.000000 CCAFS SLC 40 0 9600.000000 VAFB SLC 4E 1 2490.000000 KSC LC 39A 1 5600.000000 KSC LC 39A 0 5300.000000 KSC LC 39A 1 6104.959412 KSC LC 39A 1 6070.000000 KSC LC 39A 0 2708.000000 KSC LC 39A 1 3669.000000 KSC LC 39A 1 9600.000000 VAFB SLC 4E 1 6761.000000 KSC LC 39A 0 2910.000000 KSC LC 39A 1 475.000000 VAFB SLC 4E 1 4990.000000 KSC LC 39A 1 9600.000000 VAFB SLC 4E 1 5200.000000 KSC LC 39A 1 3700.000000 KSC LC 39A 1 2205.000000 CCAFS SLC 40 1 9600.000000 VAFB SLC 4E 1 6104.959412 CCAFS SLC 40 1 4230.000000 CCAFS SLC 40 1 6092.000000 CCAFS SLC 40 0 9600.000000 VAFB SLC 4E 0 2760.000000 CCAFS SLC 40 0 350.000000 CCAFS SLC 40 1 3750.000000 KSC LC 39A 1 5383.850000 CCAFS SLC 40 0 2410.000000 CCAFS SLC 40 0 7076.000000 CCAFS SLC 40 1 9600.000000 VAFB SLC 4E 1 5800.000000 CCAFS SLC 40 1 7060.000000 CCAFS SLC 40 1 2800.000000 VAFB SLC 4E 1 3000.000000 KSC LC 39A 1 4000.000000 VAFB SLC 4E 1 2573.000000 CCAFS SLC 40 0 4400.000000 CCAFS SLC 40 0 9600.000000 VAFB SLC 4E 1 12259.000000 KSC LC 39A 1 2482.000000 CCAFS SLC 40 1 13620.000000 CCAFS SLC 40 1 1425.000000 VAFB SLC 4E 1 2227.700000 CCAFS SLC 40 1 6500.000000 CCAFS SLC 40 0 15600.000000 CCAFS SLC 40 1 5000.000000 CCAFS SLC 40 1 6800.000000 CCAFS SLC 40 1 15400.000000 CCAFS SLC 40 1 6104.959412 KSC LC 39A 0 15600.000000 CCAFS SLC 40 1 15400.000000 CCAFS SLC 40 0 1977.000000 CCAFS SLC 40 1 15600.000000 KSC LC 39A 0 15400.000000 KSC LC 39A 1 9525.000000 KSC LC 39A 1 15400.000000 CCAFS SLC 40 1 15400.000000 CCAFS SLC 40 1 3880.000000 CCAFS SLC 40 1 6104.959412 CCAFS SLC 40 1 15400.000000 CCAFS SLC 40 1 1600.000000 CCAFS SLC 40 1 15400.000000 KSC LC 39A 1 15400.000000 KSC LC 39A 1 15400.000000 KSC LC 39A 1 15400.000000 CCAFS SLC 40 1 3681.000000 CCAFS SLC 40 1
Now if you observe Payload Vs. Launch Site scatter point chart you will find for the VAFB-SLC launchsite there are no rockets launched for heavypayload mass(greater than 10000).
TASK 3: Visualize the relationship between success rate of each orbit type¶
Next, we want to visually check if there are any relationship between success rate and orbit type.
Let's create a bar chart for the sucess rate of each orbit
In [11]:
# HINT use groupby method on Orbit column and get the mean of Class column
# Calculate the success rate for each orbit type
orbit_success_rate = df.groupby('Orbit')['Class'].mean().reset_index()
# Sort the dataframe by success rate in descending order
orbit_success_rate = orbit_success_rate.sort_values(by='Class', ascending=False)
# Create a bar plot
plt.figure(figsize=(10, 6))
sns.barplot(x='Class', y='Orbit', data=orbit_success_rate, palette='viridis')
# Set the title and labels
plt.title('Success Rate of Each Orbit Type')
plt.xlabel('Success Rate')
plt.ylabel('Orbit')
# Show the plot
plt.show()
<ipython-input-11-d7b32a3c439f>:10: FutureWarning: Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `y` variable to `hue` and set `legend=False` for the same effect. sns.barplot(x='Class', y='Orbit', data=orbit_success_rate, palette='viridis')
In [12]:
orbit_success_rate
Out[12]:
| Orbit | Class | |
|---|---|---|
| 0 | ES-L1 | 1.000000 |
| 1 | GEO | 1.000000 |
| 3 | HEO | 1.000000 |
| 9 | SSO | 1.000000 |
| 10 | VLEO | 0.857143 |
| 5 | LEO | 0.714286 |
| 6 | MEO | 0.666667 |
| 7 | PO | 0.666667 |
| 4 | ISS | 0.619048 |
| 2 | GTO | 0.518519 |
| 8 | SO | 0.000000 |
Analyze the ploted bar chart try to find which orbits have high sucess rate.
TASK 4: Visualize the relationship between FlightNumber and Orbit type¶
For each orbit, we want to see if there is any relationship between FlightNumber and Orbit type.
In [13]:
# Plot a scatter point chart with x axis to be FlightNumber and y axis to be the Orbit, and hue to be the class value
# Create the scatter plot
plt.figure(figsize=(12, 8))
sns.scatterplot(data=df, x='FlightNumber', y='Orbit', hue='Class', palette='coolwarm')
# Set the title and labels
plt.title('Flight Number vs Orbit (Colored by Class)')
plt.xlabel('Flight Number')
plt.ylabel('Orbit')
# Show the plot
plt.legend(title='Class', loc='center left', bbox_to_anchor=(1, 0.5))
plt.grid(True)
plt.tight_layout()
plt.show()
In [14]:
# Display the DataFrame
print(df[['FlightNumber', 'Orbit', 'Class']].to_string(index=False))
FlightNumber Orbit Class
1 LEO 0
2 LEO 0
3 ISS 0
4 PO 0
5 GTO 0
6 GTO 0
7 ISS 1
8 LEO 1
9 GTO 0
10 GTO 0
11 ISS 0
12 ISS 0
13 ES-L1 1
14 ISS 0
15 GTO 0
16 ISS 0
17 LEO 1
18 PO 0
19 GTO 0
20 ISS 1
21 GTO 1
22 GTO 1
23 ISS 1
24 GTO 1
25 GTO 0
26 PO 1
27 ISS 1
28 GTO 0
29 GTO 1
30 LEO 1
31 GTO 0
32 ISS 1
33 GTO 1
34 PO 1
35 GTO 0
36 ISS 1
37 SSO 1
38 LEO 1
39 PO 1
40 GTO 1
41 GTO 1
42 ISS 1
43 PO 1
44 LEO 1
45 GTO 1
46 GTO 0
47 PO 0
48 ISS 0
49 HEO 1
50 GTO 1
51 GTO 0
52 ISS 0
53 GTO 1
54 PO 1
55 GTO 1
56 GTO 1
57 SSO 1
58 GTO 1
59 SSO 1
60 ISS 0
61 MEO 0
62 PO 1
63 ISS 1
64 ISS 1
65 VLEO 1
66 SSO 1
67 ISS 1
68 GTO 0
69 VLEO 1
70 ISS 1
71 GTO 1
72 VLEO 1
73 SO 0
74 VLEO 1
75 VLEO 0
76 ISS 1
77 VLEO 0
78 VLEO 1
79 ISS 1
80 VLEO 1
81 VLEO 1
82 MEO 1
83 GEO 1
84 VLEO 1
85 SSO 1
86 VLEO 1
87 VLEO 1
88 VLEO 1
89 VLEO 1
90 MEO 1
You should see that in the LEO orbit the Success appears related to the number of flights; on the other hand, there seems to be no relationship between flight number when in GTO orbit.
TASK 5: Visualize the relationship between Payload and Orbit type¶
Similarly, we can plot the Payload vs. Orbit scatter point charts to reveal the relationship between Payload and Orbit type
In [15]:
# Display the DataFrame
print(df[['PayloadMass', 'Orbit', 'Class']].to_string(index=False))
PayloadMass Orbit Class 6104.959412 LEO 0 525.000000 LEO 0 677.000000 ISS 0 500.000000 PO 0 3170.000000 GTO 0 3325.000000 GTO 0 2296.000000 ISS 1 1316.000000 LEO 1 4535.000000 GTO 0 4428.000000 GTO 0 2216.000000 ISS 0 2395.000000 ISS 0 570.000000 ES-L1 1 1898.000000 ISS 0 4707.000000 GTO 0 2477.000000 ISS 0 2034.000000 LEO 1 553.000000 PO 0 5271.000000 GTO 0 3136.000000 ISS 1 4696.000000 GTO 1 3100.000000 GTO 1 2257.000000 ISS 1 4600.000000 GTO 1 5500.000000 GTO 0 9600.000000 PO 1 2490.000000 ISS 1 5600.000000 GTO 0 5300.000000 GTO 1 6104.959412 LEO 1 6070.000000 GTO 0 2708.000000 ISS 1 3669.000000 GTO 1 9600.000000 PO 1 6761.000000 GTO 0 2910.000000 ISS 1 475.000000 SSO 1 4990.000000 LEO 1 9600.000000 PO 1 5200.000000 GTO 1 3700.000000 GTO 1 2205.000000 ISS 1 9600.000000 PO 1 6104.959412 LEO 1 4230.000000 GTO 1 6092.000000 GTO 0 9600.000000 PO 0 2760.000000 ISS 0 350.000000 HEO 1 3750.000000 GTO 1 5383.850000 GTO 0 2410.000000 ISS 0 7076.000000 GTO 1 9600.000000 PO 1 5800.000000 GTO 1 7060.000000 GTO 1 2800.000000 SSO 1 3000.000000 GTO 1 4000.000000 SSO 1 2573.000000 ISS 0 4400.000000 MEO 0 9600.000000 PO 1 12259.000000 ISS 1 2482.000000 ISS 1 13620.000000 VLEO 1 1425.000000 SSO 1 2227.700000 ISS 1 6500.000000 GTO 0 15600.000000 VLEO 1 5000.000000 ISS 1 6800.000000 GTO 1 15400.000000 VLEO 1 6104.959412 SO 0 15600.000000 VLEO 1 15400.000000 VLEO 0 1977.000000 ISS 1 15600.000000 VLEO 0 15400.000000 VLEO 1 9525.000000 ISS 1 15400.000000 VLEO 1 15400.000000 VLEO 1 3880.000000 MEO 1 6104.959412 GEO 1 15400.000000 VLEO 1 1600.000000 SSO 1 15400.000000 VLEO 1 15400.000000 VLEO 1 15400.000000 VLEO 1 15400.000000 VLEO 1 3681.000000 MEO 1
In [16]:
# Plot a scatter point chart with x axis to be Payload and y axis to be the Orbit, and hue to be the class value
# Create the scatter plot
plt.figure(figsize=(12, 8))
sns.scatterplot(data=df, x='PayloadMass', y='Orbit', hue='Class', palette='coolwarm')
# Set the title and labels
plt.title('Payload vs Orbit (Colored by Class)')
plt.xlabel('Payload Mass (kg)')
plt.ylabel('Orbit')
# Show the plot
plt.legend(title='Class', loc='center left', bbox_to_anchor=(1, 0.5))
plt.grid(True)
plt.tight_layout()
plt.show()
With heavy payloads the successful landing or positive landing rate are more for Polar,LEO and ISS.
However for GTO we cannot distinguish this well as both positive landing rate and negative landing(unsuccessful mission) are both there here.
TASK 6: Visualize the launch success yearly trend¶
You can plot a line chart with x axis to be Year and y axis to be average success rate, to get the average launch success trend.
The function will help you get the year from the date:
In [17]:
# A function to Extract years from the date
year=[]
def Extract_year():
for i in df["Date"]:
year.append(i.split("-")[0])
return year
Extract_year()
df['Date'] = year
df.head()
Out[17]:
| FlightNumber | Date | BoosterVersion | PayloadMass | Orbit | LaunchSite | Outcome | Flights | GridFins | Reused | Legs | LandingPad | Block | ReusedCount | Serial | Longitude | Latitude | Class | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2010 | Falcon 9 | 6104.959412 | LEO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0003 | -80.577366 | 28.561857 | 0 |
| 1 | 2 | 2012 | Falcon 9 | 525.000000 | LEO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0005 | -80.577366 | 28.561857 | 0 |
| 2 | 3 | 2013 | Falcon 9 | 677.000000 | ISS | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B0007 | -80.577366 | 28.561857 | 0 |
| 3 | 4 | 2013 | Falcon 9 | 500.000000 | PO | VAFB SLC 4E | False Ocean | 1 | False | False | False | NaN | 1.0 | 0 | B1003 | -120.610829 | 34.632093 | 0 |
| 4 | 5 | 2013 | Falcon 9 | 3170.000000 | GTO | CCAFS SLC 40 | None None | 1 | False | False | False | NaN | 1.0 | 0 | B1004 | -80.577366 | 28.561857 | 0 |
In [18]:
# Plot a line chart with x axis to be the extracted year and y axis to be the success rate
# Ensure 'Date' column is in datetime format
df['Date'] = pd.to_datetime(df['Date'])
# Extract year from the 'Date' column
df['Date'] = df['Date'].dt.year
# Calculate success rate for each year
success_rate = df.groupby('Date')['Class'].apply(lambda x: (x == 1).mean() * 100)
# Plot line chart
plt.figure(figsize=(10, 6))
success_rate.plot(marker='o', color='b', linestyle='-')
# Set the title and labels
plt.title('Success Rate Over Years')
plt.xlabel('Year')
plt.ylabel('Success Rate (%)')
# Set ticks for x-axis
plt.xticks(success_rate.index)
# Show the plot
plt.grid(True)
plt.tight_layout()
plt.show()
In [23]:
success_rate
Out[23]:
Date 2010 0.000000 2012 0.000000 2013 0.000000 2014 33.333333 2015 33.333333 2016 62.500000 2017 83.333333 2018 61.111111 2019 90.000000 2020 84.210526 Name: Class, dtype: float64
you can observe that the sucess rate since 2013 kept increasing till 2020
Features Engineering¶
By now, you should obtain some preliminary insights about how each important variable would affect the success rate, we will select the features that will be used in success prediction in the future module.
In [19]:
features = df[['FlightNumber', 'PayloadMass', 'Orbit', 'LaunchSite', 'Flights', 'GridFins', 'Reused', 'Legs', 'LandingPad', 'Block', 'ReusedCount', 'Serial']]
features.head()
Out[19]:
| FlightNumber | PayloadMass | Orbit | LaunchSite | Flights | GridFins | Reused | Legs | LandingPad | Block | ReusedCount | Serial | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 6104.959412 | LEO | CCAFS SLC 40 | 1 | False | False | False | NaN | 1.0 | 0 | B0003 |
| 1 | 2 | 525.000000 | LEO | CCAFS SLC 40 | 1 | False | False | False | NaN | 1.0 | 0 | B0005 |
| 2 | 3 | 677.000000 | ISS | CCAFS SLC 40 | 1 | False | False | False | NaN | 1.0 | 0 | B0007 |
| 3 | 4 | 500.000000 | PO | VAFB SLC 4E | 1 | False | False | False | NaN | 1.0 | 0 | B1003 |
| 4 | 5 | 3170.000000 | GTO | CCAFS SLC 40 | 1 | False | False | False | NaN | 1.0 | 0 | B1004 |
TASK 7: Create dummy variables to categorical columns¶
Use the function get_dummies and features dataframe to apply OneHotEncoder to the column Orbits, LaunchSite, LandingPad, and Serial. Assign the value to the variable features_one_hot, display the results using the method head. Your result dataframe must include all features including the encoded ones.
In [20]:
# HINT: Use get_dummies() function on the categorical columns
# Columns to apply one-hot encoding
columns_to_encode = ['Orbit', 'LaunchSite', 'LandingPad', 'Serial']
# Apply one-hot encoding
features_one_hot = pd.get_dummies(features, columns=columns_to_encode)
# Display the results
print(features_one_hot.head())
FlightNumber PayloadMass Flights GridFins Reused Legs Block \ 0 1 6104.959412 1 False False False 1.0 1 2 525.000000 1 False False False 1.0 2 3 677.000000 1 False False False 1.0 3 4 500.000000 1 False False False 1.0 4 5 3170.000000 1 False False False 1.0 ReusedCount Orbit_ES-L1 Orbit_GEO ... Serial_B1048 Serial_B1049 \ 0 0 0 0 ... 0 0 1 0 0 0 ... 0 0 2 0 0 0 ... 0 0 3 0 0 0 ... 0 0 4 0 0 0 ... 0 0 Serial_B1050 Serial_B1051 Serial_B1054 Serial_B1056 Serial_B1058 \ 0 0 0 0 0 0 1 0 0 0 0 0 2 0 0 0 0 0 3 0 0 0 0 0 4 0 0 0 0 0 Serial_B1059 Serial_B1060 Serial_B1062 0 0 0 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 0 [5 rows x 80 columns]
TASK 8: Cast all numeric columns to float64¶
Now that our features_one_hot dataframe only contains numbers cast the entire dataframe to variable type float64
In [21]:
# HINT: use astype function
# Cast the entire DataFrame to float64
features_one_hot = features_one_hot.astype('float64')
# Display the results
print(features_one_hot.head())
FlightNumber PayloadMass Flights GridFins Reused Legs Block \ 0 1.0 6104.959412 1.0 0.0 0.0 0.0 1.0 1 2.0 525.000000 1.0 0.0 0.0 0.0 1.0 2 3.0 677.000000 1.0 0.0 0.0 0.0 1.0 3 4.0 500.000000 1.0 0.0 0.0 0.0 1.0 4 5.0 3170.000000 1.0 0.0 0.0 0.0 1.0 ReusedCount Orbit_ES-L1 Orbit_GEO ... Serial_B1048 Serial_B1049 \ 0 0.0 0.0 0.0 ... 0.0 0.0 1 0.0 0.0 0.0 ... 0.0 0.0 2 0.0 0.0 0.0 ... 0.0 0.0 3 0.0 0.0 0.0 ... 0.0 0.0 4 0.0 0.0 0.0 ... 0.0 0.0 Serial_B1050 Serial_B1051 Serial_B1054 Serial_B1056 Serial_B1058 \ 0 0.0 0.0 0.0 0.0 0.0 1 0.0 0.0 0.0 0.0 0.0 2 0.0 0.0 0.0 0.0 0.0 3 0.0 0.0 0.0 0.0 0.0 4 0.0 0.0 0.0 0.0 0.0 Serial_B1059 Serial_B1060 Serial_B1062 0 0.0 0.0 0.0 1 0.0 0.0 0.0 2 0.0 0.0 0.0 3 0.0 0.0 0.0 4 0.0 0.0 0.0 [5 rows x 80 columns]
We can now export it to a CSV for the next section,but to make the answers consistent, in the next lab we will provide data in a pre-selected date range.
In [22]:
features_one_hot.to_csv('dataset_part_3.csv', index=False)
features_one_hot.to_csv('dataset_part_3.csv', index=False)
Authors¶
Change Log¶
| Date (YYYY-MM-DD) | Version | Changed By | Change Description |
|---|---|---|---|
| 2022-11-09 | 1.0 | Pratiksha Verma | Converted initial version to Jupyterlite |