Winning Space Race

with Data Science

By Veronika Kozia
14/11/2024
OUTLINE
• Executive summary
• Introduction
• Methodology
• Results
• Conclusion
• Appendix
 EXECUTIVE SUMMARY
• SpaceX has made significant strides in
the aerospace industry by successfully
reusing the first stage of its rockets.
This innovation has drastically reduced
launch costs and improved efficiency.
With its commitment to sustainable
space travel, SpaceX is set to continue
using this groundbreaking technology,
ensuring ongoing success and leading
the way in making space more
accessible.
INTRODUCTION
• Unlike other rocket providers, SpaceX's Falcon 9 can recover the first stage. Sometimes the first
stage does not land, sometimes it will crash. Other times, Space X will sacrifice the first stage due
to the mission parameters like payload, orbit, and customer.

• Space Y that would like to compete with SpaceX founded by Billionaire industrialist
Allon Musk. Our job is to determine the price of each launch. We will do this by gathering
information about Space X and creating dashboards for our team. We will also determine if SpaceX
will reuse the first stage. Instead of using rocket science to determine if the first stage will land
successfully, we will train a machine learning model and use public information to predict if
SpaceX will reuse the first stage.
 The nature of this analysis is to evaluate SpaceX's
reusable rocket technology, specifically focusing on
the first stage reuse during launches. By examining
WHAT WE WILL LEARN:
launch data, success rates, and technical
parameters, we aim to assess the effectiveness and
sustainability of this innovative approach.

The key questions guiding this analysis are: What is

the success rate of reusing the first stage of
The problem at hand is determining how well the rockets? How does this reuse impact overall cost
reuse of the first stage aligns with SpaceX's goals of and efficiency? What distances do the first stages
reducing launch costs and increasing the frequency cover post-launch, and how does this affect
and reliability of space missions. This requires a recovery and refurbishment? Which launch sites
detailed analysis of various factors that contribute and conditions yield higher success rates? These
to or hinder success. questions will provide insights into the performance
and future potential of SpaceX's reusable rocket
technology.
Section 1:
METHODOLOGY
Methodology
Data resources:
Public Datasets: Data from NASA, ESA, and other space
agencies on global launch activities and environmental
conditions.

• Perform exploratory data analysis(EDA) using visualization

and SQL
• Perform predictive analysis using classification models
DATA COLLECTION
• Plan for Collecting Data:
• Gather detailed launch data from SpaceX's official records.
• Utilize public datasets from space agencies for additional context.
• Review relevant scientific publications for technical insights.
• Analyze industry reports for broader trends and challenges.
DATA COLLECTION – SPACEX API &
SCRAPING
• Space-Y/jupyter-labs-spacex-
data-collection-api.ipynb at
main · verokozia/Space-Y
• The table of the rocket
launches, outcomes, dates,
and all the important details of
each launch – all to analyze the
success of the technology
used.
Data Collection - scraping
• Space-Y/jupyter-labs-
webscraping.ipynb at main ·
verokozia/Space-Y
Data wrangling
• Determined the number of
launches from each site, as
well as number of launches to
each orbit and outcomes.
Adjusted the table.
• Space-Y/labs-jupyter-spacex-
Data wrangling.ipynb at main ·
verokozia/Space-Y
EDA with Data
Visualization
Used different charts and graphs to show
the relation between the variables to
enhance the deeper understanding of the
content and its progression. (see more
charts by the link mentioned)

Space-Y/edadataviz.ipynb at main ·
verokozia/Space-Y
EDA with SQL
Used to analyze, compare and gazer
insights from the dataset, such as number
of different mission outcomes (see more
analysis by the link below)

Space-Y/jupyter-labs-eda-sql-
coursera_sqllite.ipynb at main ·
verokozia/Space-Y
Predictive analysis
(classification)
Used to train models to predict the
possible outcomes and relations. For
instance, the chance of the rocket to land
successfully. (see more examples by the
link below)

Space-Y/SpaceX-Machine-Learning-
Prediction-Part-5-v1 (1).ipynb at main ·
verokozia/Space-Y
Results

• The analysis confirms that SpaceX's reuse of the first stage is

both effective and sustainable. The success rate of reusing
the first stage has been impressively high, with the majority of
missions achieving successful landings and subsequent
reuses. This reuse significantly reduces overall launch costs
and increases efficiency by minimizing the need for new
production. Post-launch, the first stages typically cover
distances that allow for efficient recovery and refurbishment,
often landing on designated drone ships or landing pads.
Launch sites like Cape Canaveral and Vandenberg, with
favorable geographic and weather conditions, have
demonstrated higher success rates in reusability, highlighting
the importance of location in the success of these missions.
Overall, the reuse strategy not only meets SpaceX's goals of
cost reduction and increased launch frequency but also sets
a new standard in the aerospace industry.
SECTION 2
Insights drawn from EDA
Launch Site vs Flight Number
This gives an insight that VAFB SLC 4E sight
only had flights with numbers up to ~70,
Whereas KSC LC 39A starts from ~25
And CCAFS SLC 40 does not have any
flights in between numbers ~25 to 40.
Launch Site vs Pay
Load Mass

• Now if you observe Payload

Mass 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).
Orbit type vs success
rate
Orbits ES-L1, GEO, HEO and SSO
have the highest success rates,
whereas SO orbit shows no
successful attempts of lauch.
Flight number vs Orbit
You can observe that in the
LEO orbit, success seems to be
related to the number of
flights. Conversely, in the GTO
orbit, there appears to be no
relationship between flight
number and success.
Payload mass vs Orbit
With heavy payloads the
successful landing or positive
landing rate are more for
Polar,LEO and ISS.
However, for GTO, it's difficult
to distinguish between
successful and unsuccessful
landings as both outcomes are
present.
Launch success yearly
trend
The success rate has been
increasing since 2013, with a
little throwback in 2018 and
2020. This indicates success
of the technology of the first
stage.
Launch sites
There are three main
launch sites:
CCAFS LC-40
VAFB SLC-4E
KSC LC-39A
Launch cite names
begin with 'CCA' (5)
The only launch site
that begins with 'CCA'
is CCAFS LC-40. It has
been used by both
Nasa and our
competitor, SpaceX. All
the mission outcomes
are successful.
Total payload mass
• Lift Capability: It indicates how much
weight the rocket can carry into space,
which is critical for understanding the
vehicle’s performance and suitability
for different types of missions, whether
launching satellites, space probes, or
crewed missions.
• Mission Scope: The payload mass can
determine the scope and complexity of
the mission. Heavier payloads might
include larger satellites or multiple
payloads, which require more powerful
rockets and fuel efficiency.
Average payload mass
for F9 v1.1
Gives an insight about its
capabilities, cost and
efficiency.
These figures highlight the
rocket's capability to carry
significant payloads, making
it a versatile and efficient
choice for a variety of
missions.
First successful
landing
The first successful landing
took place in 2015,
December 22nd.
Successful drone ship
landing with payload
between 4000 and 6000.
Quite a few missions went
successful with this
technology, which signifies its
efficiency and relative
reliability.
Total number of
successful and failure
mission outcomes.
Most missions were
considered successful,
which indicates company's
diligent approach towards
their goals. The first stage
seems to be quite useful and
thus is to be used again.
Boosters carried
maximum payload
The list of boosters that have
carried the maximum
payload, indicating the most
powerful boosters of all used.
Valuable information for the
future launches.
2015 launch records
The month, landing outcome,
booster version and launch
site for all the failed (drone
ship) missions in 2015.
Rank landing outcomes
between 2010/06/04
and 2017/03/20
The rank of the different
outcomes through seven
years. Mostly there are no
attempts (10) and the most
practiced ones are with drone
ships with success rate of
50%. Only one drone ship
mission is precluded.
Section 3
Predictive analysis and classification
(other sections are missing because the labs did not go well on my PC. My
apologies. I hope for your understanding.)
Classification
accuracy
The results below show us that all the models perform pretty well, although the
decision tree sems to be less accurate than the other methods that maintain
the same accuracy.
The confusion matrix
The confusion matrix for the
models that show the best
performance accuracy
(82.3333%).
The confusion matrix is the
same for all the models that
have the same performance
accuracy.
CONCLUSIONS
• SpaceX has demonstrated remarkable success with its reusable rocket
technology, specifically the first stage reuse.
• With high success rates in landing and reusing the first stage, the
company has significantly reduced launch costs and improved
efficiency.
• By analyzing data from various missions, it is evident that the reuse of
the first stage not only meets SpaceX's objectives of cost reduction but
also enhances the sustainability of space travel.
• Launch sites like CCA and VSC have shown higher success rates,
emphasizing the importance of location in mission success.
• Overall, SpaceX's innovative approach has set a new standard in the
aerospace industry, promising a future of more affordable and frequent
space missions.
Points to discuss:

• Cost Efficiency:
In what ways has SpaceX's reuse of the first stage
reduced launch costs compared to traditional single-
use rockets?
How could these cost savings impact the future of
space exploration and commercialization?
• Future Prospects:
What future missions and goals can we expect from
SpaceX as they continue to innovate in reusable rocket
technology?
How might these advancements accelerate the
timeline for missions to Mars and beyond?
Appendix
All the further information about the code, queries, charts and
tables can be reached by the link below:

verokozia/Space-Y
Thank you.