SlideShare a Scribd company logo
Decoding SpaceX
Unveiling Launch Success with Data
Science
Mariano G. Tinti
2
• Executive Summary
• Introduction
• Methodology
• Results
• Conclusion
• Appendix
Outline
3
METHODOLOGIES
• Collect data using SpaceX REST API and web scraping techniques​
• Wrangle data to create success/fail outcome variable​
• Explore data with data visualization techniques, considering the following factors: payload, launch site, flight
number and yearly trend​
• Analyze the data with SQL, calculating the following statistics: total payload, payload range for successful
launches, and total # of successful and failed outcomes​
• Explore launch site success rates and proximity to geographical markers​
• Visualize the launch sites with the most success and successful payload ranges​
• Build Models to predict landing outcomes using logistic regression, support vector machine (SVM), decision
tree and K-nearest neighbor (KNN)
Executive Summary
4
Introduction
• Space Y is a new and ambitious company that aims to compete with
SpaceX, the leading private space company in the world.
• SpaceX was founded in 2002 by Elon Musk, with the ultimate goal of
enabling human life on Mars. It has achieved many remarkable milestones
in the history of space exploration.
Based on public information provided by SpaceX we aim to:
• Determine if a landing will be successful
• Know how different variables affect the outcome of the landing, such as
launch site, payload mass, booster version, etc.
• Find correlations between launch sites and success rates
• Describe some characteristics of the launch sites
5
Section 1
6
Executive Summary
• Data collection methodology:
• Scraping from the Space X API and Wikipedia
• Perform data wrangling
• Collected data was enriched by creating a landing outcome label based on outcome
data after summarizing and analyzing features.
• Perform exploratory data analysis (EDA) using visualization and SQL
• Perform interactive visual analytics using Folium and Plotly Dash
• Perform predictive analysis using classification models
• Data was standardized and divided into TEST and TRAINING sets. Four models were
proposed and the most successful was selected based on the highest accuracy
achieved.
Methodology
7
• SpaceX offers a public API from
where data can be obtained.
• https://api.spacexdata.com/v4/rock
ets/
• GitHub URL of the completed
SpaceX API calls notebook:
https://github.com/marianotinti/AppliedDataS
cienceCapstone/blob/main/jupyter-labs-
spacex-data-collection-api.ipynb
Data Collection – SpaceX API
Convert
new
dataframe
to CSV
Extract
relevant
data to
dataframe
Read .json
response
and convert
into a
dataframe
Request
API and
parse the
SpaceX
launch data
8
• More data from SpaceX
launches was scraped from
Wikipedia
GitHub URL of the completed
SpaceX API calls notebook:
https://github.com/marianotinti/AppliedDa
taScienceCapstone/blob/main/jupyter-
labs-webscraping.ipynb
Data Collection - Scraping
Convert
new
dataframe
to CSV
Extract
relevant
data to
dataframe
Create
beautiful
soup object
from the
response
Request
Falcon9
html page
via HTTP
GET
9
• Exploratory data analysis (EDA) was performed on the dataset.
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/labs-jupyter-
spacex-Data%20wrangling.ipynb
Data Wrangling
Load the data
Explore and find
patterns
Summarize
Create the labels of
mission outcomes
10
• To explore data, scatterplots, barplots and line charts were used to
visualize the relationship between pairs of features such as:
• Payload Mass X Flight Number, Launch Site X Flight Number, Launch Site X
Payload Mass, orbit and Flight Number, Payload and Orbit.
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/IBM-DS0321EN-
SkillsNetwork_labs_module_2_jupyter-labs-eda-dataviz.ipynb.jupyterlite.ipynb
EDA with Data Visualization
11
• To better understand SpaceX data set, following SQL queries/operations were performed on an IBM
DB2 cloud instance:
1. Display the names of the unique launch sites in the space mission
2. Display 5 records where launch sites begin with the string 'CCA’
3. Display the total payload mass carried by boosters launched by NASA (CRS)
4. Display average payload mass carried by booster version F9 v1.1
5. List the date when the first successful landing outcome in ground pad was achieved.
6. List the names of the boosters which have success in drone ship and have payload mass greater than 4000
but less than 6000
7. List the total number of successful and failure mission outcomes
8. List the names of the booster_versions which have carried the maximum payload mass. Use a subquery
9. List the failed landing_outcomes in drone ship, their booster versions, and launch site names for in year 2015
10. Rank the count of landing outcomes (such as Failure (drone ship) or Success (ground pad) between the date
2010-06- 04 and 2017-03-20, in descending order
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/jupyter-labs-eda-sql-
coursera_sqllite.ipynb
EDA with SQL
12
We took advantage of geospatial data with the use of folium, which also allowed us to perform interactive visual analysis.
• Launch sites were marked on the map with marker objects.
• Added ‘folium.circle’ and ‘folium.marker’ to highlight circle area with a text label over each launch site.
• Added a ‘MarkerCluster()’ to show launch success (green) and failure (red) markers for each launch site.
• Calculated distances between a launch site to its proximities (e.g., coastline, railroad, highway, city)
• Added ‘MousePosition() to get coordinate for a mouse position over a point on the map
• Added ‘folium.Marker()’ to display distance (in KM) on the point on the map (e.g., coastline, railroad, highway, city)
• Added ‘folium.Polyline()’ to draw a line between the point on the map and the launch site
• Repeated steps above to add markers and draw lines between launch sites and proximities – coastline, railroad, highway,
city
• This way we could understand that launch sites are in close proximity to: HIGHWAYS, RAILWAYS, COASTLINES AND
CITIES
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/lab_jupyter_launch_site_locati
on.ipynb
Build an Interactive Map with Folium
13
1. Added a Launch Site Drop-down Input component to the dashboard to provide an
ability to filter Dashboard visual by all launch sites or a particular launch site
2. Added a Pie Chart to the Dashboard to show total success launches when ‘All
Sites’ is selected and show success and failed counts when a particular site is
selected
3. Added a Payload range slider to the Dashboard to easily select different payload
ranges to identify visual patterns
4. Added a Scatter chart to observe how payload may be correlated with mission
outcomes for selected site(s). The color-label Booster version on each scatter point
provided missions outcomes with different boosters
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/Dash.py
Build a Dashboard with Plotly Dash
14
• Four classification models were compared: logistic regression, support
vector machine and decision tree.
• https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/IBM-DS0321EN-
SkillsNetwork_labs_module_4_SpaceX_Machine_Learning_Prediction_Part_5.jupyterlite.ipy
nb
Predictive Analysis (Classification)
Load the dataset
and create
NumPy arrays
Standardize data
Split into TRAIN
and TEST
datasets
Create models
(Log Regression,
Decision tree, K-
Nearest
Neighbors nad
Support Vector
Machine
Test each model
with
combinations of
hyperparameters
Calculate
Accuracy of each
model and make
a decision
• Exploratory data analysis results
• Interactive analytics demo in screenshots
• Predictive analysis results
• Accuracy for Logistics Regression method: 0.8333333333333334
• Accuracy for Support Vector Machine method: 0.8333333333333334
• Accuracy for Decision tree method: 0.8888888888888888
• Accuracy for K nearest neighbors method: 0.8333333333333334 15
Results
Section 2
17
• CCAF5 SLC 40 is where most of recent launches were successful
• Success rates (Class=1) increases as the number of flights increases
Flight Number vs. Launch Site
18
• For launch site ‘VAFB SLC 4E’,
there are no rockets launched for
payload greater than 10,000 kg.
• Payloads over 12,000kg seems to
be possible only on CCAFS SLC
40 and KSC LC 39A launch sites.
• Percentage of successful launch
(Class=1) increases for launch
site ‘VAFB SLC 4E’ as the
payload mass increases
• There is no clear correlation or
pattern between launch site and
payload mass
Payload vs. Launch Site
19
• The biggest success rates happens to
orbits:
• ES-L1;
• GEO;
• HEO;
• SSO.
• Followed by:
• VLEO (above 80%)
• LFO (above 70%).
GTO orbit has the lowest success rate.
SO orbit has no successful launches
Success Rate vs. Orbit Type
20
• For orbit VLEO, first successful
landing (class=1) doesn’t occur until
60+ number of flights
• For most orbits (LEO, ISS, PO,
SSO, MEO, VLEO) successful
landing rates appear to increase
with flight numbers
• There is no relationship between
flight number and orbit for GTO
Flight Number vs. Orbit Type
21
• Successful landing rates
(Class=1) appear to
increase with pay load for
orbits LEO, ISS, PO, and
SSO
• For GEO orbit, there is not
clear pattern between
payload and successful
launches
Payload vs. Orbit Type
22
✔ Success rates improved over time
o Success rate (Class=1) increased
by about 80% between 2013 and
2020
o Success rates remained the same
between 2010 and 2013 and
between 2014 and 2015
o Success rates decreased between
2017 and 2018 and between 2019
and 2020
Launch Success Yearly Trend
23
• Query:
%sql Select distinct Launch_site from spacextbl
‘distinct’ returns only unique values
• Result:
All Launch Site Names
24
• Query:
%sql select * from spacextbl where Launch_Site LIKE ‘CCA%’ limit 5;
Keyword LIKE and format ‘CCA%’ returns records in the Launch Site column
starting with “CCA”
Limit 5 limits the number of records shown to 5
• Result:
Launch Site Names Begin with 'CCA'
25
• Query:
%sql select sum(PAYLOAD_MASS_KG_) from spacextbl where Customer = ‘NASA
(CRS)’
Sum adds the results of the mentioned column and returns the total payload mass for the
described customer
• Result:
Total Payload Mass
26
• Query:
%sql select avg (PAYLOAD_MASS_KG_) from spacextbl where Booster_Version LIKE
‘F9 v1.1’
‘avg’ keyword returns the average of payload mass in ‘PAYLOAD_MASS_KG’ column
where
booster version is ‘F9 v1.1’
• Result:
Average Payload Mass by F9 v1.1
27
• Query:
%sql SELECT min(DATE) FROM SPACEXTBL WHERE
LANDING__OUTCOME='Success (ground pad)’
‘min(Date)’ selects the first or the oldest date from the ‘Date’ column where first
successful landing on group pad was achieved
Where clause defines the criteria to return date for scenarios where ‘Landing_Outcome’
value is equal to ‘Success (ground pad)’
• Result:
First Successful Ground Landing Date
28
• Query:
%sql SELECT BOOSTER_VERSION FROM SPACEXTBL WHERE
PAYLOAD_MASS__KG_ between 4000 and 6000 AND LANDING__OUTCOME=
'Success (drone ship)’ ;
The query finds the booster version where payload mass is greater than 4000 but less
than 6000 and the landing outcome is success in drone ship
The ‘and’ operator in the where clause returns booster versions where both conditions in
the where clause are true
• Result:
Successful Drone Ship Landing with Payload between 4000 and 6000
29
• Query:
%sql SELECT COUNT(*) FROM SPACEXTBL WHERE
MISSION_OUTCOME LIKE '%Success%' OR MISSION_OUTCOME
LIKE '%Failure%’
The ‘group by’ keyword arranges identical data in a column in to group
In this case, number of mission outcomes by types of outcomes are grouped in column
‘counts’
• Result:
Total Number of Successful and Failure Mission Outcomes
30
• Query:
%sql SELECT BOOSTER_VERSION FROM SPACEXTBL
WHERE PAYLOAD_MASS__KG_ = (SELECT
MAX(PAYLOAD_MASS__KG_) FROM SPACEXTBL)
The sub query returns the maximum payload mass by using keywork ‘max’ on the pay
load
mass column
The main query returns booster versions and respective payload mass where payload
mass is maximum with value of 15600
• Result:
Boosters Carried Maximum Payload
31
• Query:
%sql SELECT TO_CHAR(TO_DATE(MONTH("DATE"), 'MM'), 'MONTH') AS MONTH_NAME, 
LANDING__OUTCOME AS LANDING__OUTCOME, 
BOOSTER_VERSION AS BOOSTER_VERSION, 
LAUNCH_SITE AS LAUNCH_SITE 
FROM SPACEXTBL WHERE LANDING__OUTCOME = 'Failure (drone ship)' AND "DATE" LIKE '%2015%’
The query lists landing outcome, booster version, and the launch site where landing outcome is failed in drone ship and the year is 2015
The ‘and’ operator in the where clause returns booster versions where both conditions in the
where clause are true
The ‘year’ keywork extracts the year from column ‘Date
The results identify launch site as ‘CCAFS LC-40’ and booster version as F9 v1.1 B1012 and B1015 that had failed landing outcomes in drop ship in
the year 2015
• Result:
2015 Launch Records
32
• Query:
%sql SELECT "DATE", COUNT(LANDING__OUTCOME) as COUNT FROM SPACEXTBL 
WHERE "DATE" BETWEEN '2010-06-04' and '2017-03-20' AND LANDING__OUTCOME LIKE
'%Success%' 
GROUP BY "DATE" 
ORDER BY COUNT(LANDING__OUTCOME) DESC
The ‘group by’ key word arranges data in column ‘Landing__Outcome’ into groups
The ‘between’ and ‘and’ keywords return data that is between 2010-06-04 and 2017-03-20
The ‘order by’ keyword arranges the counts column in descending order
The result of the query is a ranked list of landing outcome counts per the specified date range
• Result:
Rank Landing Outcomes Between 2010-06-04 and 2017-03-20
Section 3
34
Falcon 9 Launch Sites Map
Fig 1 – Launch
Sites
Fig 3 – Launch sites located in Florida
Fig 2 – VAFB SLC 4E Launch site
Figure 1 on left displays the map with Falcon 9 launch sites that
are located in the United States (in California and Florida). Each
launch site contains a circle, label, and a popup to highlight the
location, name and number of launches that happened on the
launch site. It is also evident that all launch sites are near the
coast.
Figure 2 and Figure 3 zoom in to the launch
sites to display 4 launch sites:
• VAFB SLC-4E (CA)
• CCAFS LC-40 (FL)
• KSC LC-39A (FL)
• CCAFS SLC-40 (FL)
35
Falcon9 Success/Failed Launch Map for all Launch
Sites
Fig 4 – VAFB SLC-4E Launch
Site with success/failed
markers
Fig 3 – KSC LC -39A Launch
Site with success/failed
markers
Fig 2 – CCAFS LC-40 Launch
Site with success/failed
markers
Fig 1 – CCAFS SLC-40
Launch Site with
success/failed markers
Figure 1 Through 4 zoom in to the
launch sites to display the
failed/successful launches
• KSC LC-39A has the greatest
number of successful launches
36
Falcon9 – Launch Site to proximity Distance Map
Fig 1 – VAFB SLC-4E Launch Site with calculated distances to:
Railways, Highways and the coastline
Fig 2 – VAFB SLC-4E Launch Site with calculated distances to the nearest city
Figure 1 provides a zoom in view for the VAFB launch site, and shows other proximities such a coastline, railroad, and highway
with respective distances from the Launch Site
Figure 2 displays the proximity to the City Lompoc, which is located further away from Launch Site compared to other proximities
such as coastline, railroad, highway, etc. The map also displays a marker with city distance from the Launch Site (14.09 km)
In general, cities are located away from the Launch Sites to minimize impacts of any accidental impacts to the general public and
infrastructure. Launch Sites are strategically located near the coastline, railroad, and highways to provide easy access to
resources
Section 4
38
Launch Success Counts For All Sites
• Launch Site ‘KSC LC-39A’ has the highest launch success rate
• Launch Site ‘CCAFS SLC40’ has the lowest launch success rate
39
• KSC LC-39A Launch Site has the highest launch success rate and count
• Launch success rate is 76.9%
• Launch success failure rate is 23.1%
Launch Site with Highest Launch Success Ratio
40
• Most successful launches are in the payload range from 2000 to about 5500
• Booster version category ‘FT’ has the most successful launches
• Only booster with a success launch when payload is greater than 6k is ‘B4’
Payload vs. Launch Outcome Scatter Plot for All Sites
Section 5
42
Classification Accuracy
Fig 1 – Bar plot showing the
methods employe d and the
accuracies achieved.
43
Confusion Matrix
Fig 1 – Confusion matrix of the
Decision tree model.
Per the confusion matrix, the
classifier made 18 predictions
This shows that the model has a
slight problem interpreting false
negatives.
The previous models performed
worse and had equal problems
with false negatives and positives.
This model is more SPECIFIC
towards knowing what makes the
rockets NOT LAND an 88% of
times
44
• As the numbers of flights increase, the first stage is more likely to land successfully
• Success rates appear go up as Payload increases but there is no clear correlation between
• Payload mass and success rates
• Launch success rate increased by about 80% from 2013 to 2020
• Launch Site ‘KSC LC-39A’ has the highest launch success rate and Launch Site ‘CCAFS
SLC40’ has the lowest launch success rate
• Orbits ES-L1, GEO, HEO, and SSO have the highest launch success rates and orbit GTO the
lowest
• Launch sites are located strategically away from the cities and closer to coastline, railroads,
and highways
• The best performing Machine Learning Classfication Model is the Decision Tree with an
accuracy of about 88%.
Conclusions
45
• https://github.com/marianotinti/AppliedDataScienceCapstone/tree/main
Appendix
Mariano G Tinti - Decoding SpaceX

More Related Content

Similar to Mariano G Tinti - Decoding SpaceX

ds-capstone-template-coursera.pptx
ds-capstone-template-coursera.pptxds-capstone-template-coursera.pptx
ds-capstone-template-coursera.pptx
MustaquimAhmad1
 
Integrating PostGIS in Web Applications
Integrating PostGIS in Web ApplicationsIntegrating PostGIS in Web Applications
Integrating PostGIS in Web Applications
Command Prompt., Inc
 
NASA_CCO_status-2013 update
NASA_CCO_status-2013 updateNASA_CCO_status-2013 update
NASA_CCO_status-2013 update
Dmitry Tseitlin
 
Big Linked Data Federation - ExtremeEarth Open Workshop
Big Linked Data Federation - ExtremeEarth Open WorkshopBig Linked Data Federation - ExtremeEarth Open Workshop
Big Linked Data Federation - ExtremeEarth Open Workshop
ExtremeEarth
 
ESWC 2009 In-Use Track: SCOVO
ESWC 2009 In-Use Track: SCOVOESWC 2009 In-Use Track: SCOVO
ESWC 2009 In-Use Track: SCOVO
Michael Hausenblas
 
Patch Maps
Patch MapsPatch Maps
Patch Maps
aubreyholland
 
What are customers building with new Bing Maps capabilities
What are customers building with new Bing Maps capabilitiesWhat are customers building with new Bing Maps capabilities
What are customers building with new Bing Maps capabilities
Microsoft Tech Community
 
DepthInsight Sdk introduction
DepthInsight Sdk introductionDepthInsight Sdk introduction
DepthInsight Sdk introduction
Bo Sun
 
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
Jim Czuprynski
 
Yang enhance-voyager-user-innovations
Yang enhance-voyager-user-innovationsYang enhance-voyager-user-innovations
Yang enhance-voyager-user-innovations
ENUG
 
Open GeoData, Open GeoTools: An Introduction
Open GeoData, Open GeoTools: An IntroductionOpen GeoData, Open GeoTools: An Introduction
Open GeoData, Open GeoTools: An Introduction
Richard Cantwell
 
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
Ivan Ermilov
 
Collecting and Analysing Spaceborn Data
Collecting and Analysing Spaceborn DataCollecting and Analysing Spaceborn Data
Collecting and Analysing Spaceborn Data
Ivo Andreev
 
Commercial Crew Program status-update-1-9-13
Commercial Crew Program status-update-1-9-13Commercial Crew Program status-update-1-9-13
Commercial Crew Program status-update-1-9-13
Dmitry Tseitlin
 
Maps tek11
Maps tek11Maps tek11
Maps tek11
mayraRojo
 
Sparkling Water 5 28-14
Sparkling Water 5 28-14Sparkling Water 5 28-14
Sparkling Water 5 28-14
Sri Ambati
 
Collecting and Analysing Satellite Data with Azure Orbital
Collecting and Analysing Satellite Data with Azure OrbitalCollecting and Analysing Satellite Data with Azure Orbital
Collecting and Analysing Satellite Data with Azure Orbital
Ivo Andreev
 
LocationTech Projects
LocationTech ProjectsLocationTech Projects
LocationTech Projects
Jody Garnett
 
CARTO ENGINE
CARTO ENGINECARTO ENGINE
CARTO ENGINE
Jorge Sanz
 
chris guidice RESUME Ver2
chris guidice RESUME Ver2chris guidice RESUME Ver2
chris guidice RESUME Ver2
Christopher Guidice
 

Similar to Mariano G Tinti - Decoding SpaceX (20)

ds-capstone-template-coursera.pptx
ds-capstone-template-coursera.pptxds-capstone-template-coursera.pptx
ds-capstone-template-coursera.pptx
 
Integrating PostGIS in Web Applications
Integrating PostGIS in Web ApplicationsIntegrating PostGIS in Web Applications
Integrating PostGIS in Web Applications
 
NASA_CCO_status-2013 update
NASA_CCO_status-2013 updateNASA_CCO_status-2013 update
NASA_CCO_status-2013 update
 
Big Linked Data Federation - ExtremeEarth Open Workshop
Big Linked Data Federation - ExtremeEarth Open WorkshopBig Linked Data Federation - ExtremeEarth Open Workshop
Big Linked Data Federation - ExtremeEarth Open Workshop
 
ESWC 2009 In-Use Track: SCOVO
ESWC 2009 In-Use Track: SCOVOESWC 2009 In-Use Track: SCOVO
ESWC 2009 In-Use Track: SCOVO
 
Patch Maps
Patch MapsPatch Maps
Patch Maps
 
What are customers building with new Bing Maps capabilities
What are customers building with new Bing Maps capabilitiesWhat are customers building with new Bing Maps capabilities
What are customers building with new Bing Maps capabilities
 
DepthInsight Sdk introduction
DepthInsight Sdk introductionDepthInsight Sdk introduction
DepthInsight Sdk introduction
 
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
Politics Ain’t Beanbag: Using APEX, ML, and GeoCoding In a Modern Election Ca...
 
Yang enhance-voyager-user-innovations
Yang enhance-voyager-user-innovationsYang enhance-voyager-user-innovations
Yang enhance-voyager-user-innovations
 
Open GeoData, Open GeoTools: An Introduction
Open GeoData, Open GeoTools: An IntroductionOpen GeoData, Open GeoTools: An Introduction
Open GeoData, Open GeoTools: An Introduction
 
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
Lodstats: The Data Web Census Dataset. Kobe, Japan, 2016
 
Collecting and Analysing Spaceborn Data
Collecting and Analysing Spaceborn DataCollecting and Analysing Spaceborn Data
Collecting and Analysing Spaceborn Data
 
Commercial Crew Program status-update-1-9-13
Commercial Crew Program status-update-1-9-13Commercial Crew Program status-update-1-9-13
Commercial Crew Program status-update-1-9-13
 
Maps tek11
Maps tek11Maps tek11
Maps tek11
 
Sparkling Water 5 28-14
Sparkling Water 5 28-14Sparkling Water 5 28-14
Sparkling Water 5 28-14
 
Collecting and Analysing Satellite Data with Azure Orbital
Collecting and Analysing Satellite Data with Azure OrbitalCollecting and Analysing Satellite Data with Azure Orbital
Collecting and Analysing Satellite Data with Azure Orbital
 
LocationTech Projects
LocationTech ProjectsLocationTech Projects
LocationTech Projects
 
CARTO ENGINE
CARTO ENGINECARTO ENGINE
CARTO ENGINE
 
chris guidice RESUME Ver2
chris guidice RESUME Ver2chris guidice RESUME Ver2
chris guidice RESUME Ver2
 

Recently uploaded

Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
OnBoard
 
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
AimanAthambawa1
 
Smart Mobility Market:Revolutionizing Transportation.pdf
Smart Mobility Market:Revolutionizing Transportation.pdfSmart Mobility Market:Revolutionizing Transportation.pdf
Smart Mobility Market:Revolutionizing Transportation.pdf
Market.us
 
Improving Learning Content Efficiency with Reusable Learning Content
Improving Learning Content Efficiency with Reusable Learning ContentImproving Learning Content Efficiency with Reusable Learning Content
Improving Learning Content Efficiency with Reusable Learning Content
Enterprise Knowledge
 
Intel Unveils Core Ultra 200V Lunar chip .pdf
Intel Unveils Core Ultra 200V Lunar chip .pdfIntel Unveils Core Ultra 200V Lunar chip .pdf
Intel Unveils Core Ultra 200V Lunar chip .pdf
Tech Guru
 
kk vathada _digital transformation frameworks_2024.pdf
kk vathada _digital transformation frameworks_2024.pdfkk vathada _digital transformation frameworks_2024.pdf
kk vathada _digital transformation frameworks_2024.pdf
KIRAN KV
 
Required Documents for ISO 17021 Certification.PPT
Required Documents for ISO 17021 Certification.PPTRequired Documents for ISO 17021 Certification.PPT
Required Documents for ISO 17021 Certification.PPT
mithun772
 
How UiPath Discovery Suite supports identification of Agentic Process Automat...
How UiPath Discovery Suite supports identification of Agentic Process Automat...How UiPath Discovery Suite supports identification of Agentic Process Automat...
How UiPath Discovery Suite supports identification of Agentic Process Automat...
DianaGray10
 
Camunda Chapter NY Meetup July 2024.pptx
Camunda Chapter NY Meetup July 2024.pptxCamunda Chapter NY Meetup July 2024.pptx
Camunda Chapter NY Meetup July 2024.pptx
ZachWylie3
 
Retrieval Augmented Generation Evaluation with Ragas
Retrieval Augmented Generation Evaluation with RagasRetrieval Augmented Generation Evaluation with Ragas
Retrieval Augmented Generation Evaluation with Ragas
Zilliz
 
Mule Experience Hub and Release Channel with Java 17
Mule Experience Hub and Release Channel with Java 17Mule Experience Hub and Release Channel with Java 17
Mule Experience Hub and Release Channel with Java 17
Bhajan Mehta
 
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
FIDO Alliance
 
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
bellared2
 
Generative AI Reasoning Tech Talk - July 2024
Generative AI Reasoning Tech Talk - July 2024Generative AI Reasoning Tech Talk - July 2024
Generative AI Reasoning Tech Talk - July 2024
siddu769252
 
Zaitechno Handheld Raman Spectrometer.pdf
Zaitechno Handheld Raman Spectrometer.pdfZaitechno Handheld Raman Spectrometer.pdf
Zaitechno Handheld Raman Spectrometer.pdf
AmandaCheung15
 
Latest Tech Trends Series 2024 By EY India
Latest Tech Trends Series 2024 By EY IndiaLatest Tech Trends Series 2024 By EY India
Latest Tech Trends Series 2024 By EY India
EYIndia1
 
UX Webinar Series: Aligning Authentication Experiences with Business Goals
UX Webinar Series: Aligning Authentication Experiences with Business GoalsUX Webinar Series: Aligning Authentication Experiences with Business Goals
UX Webinar Series: Aligning Authentication Experiences with Business Goals
FIDO Alliance
 
History and Introduction for Generative AI ( GenAI )
History and Introduction for Generative AI ( GenAI )History and Introduction for Generative AI ( GenAI )
History and Introduction for Generative AI ( GenAI )
Badri_Bady
 
Uncharted Together- Navigating AI's New Frontiers in Libraries
Uncharted Together- Navigating AI's New Frontiers in LibrariesUncharted Together- Navigating AI's New Frontiers in Libraries
Uncharted Together- Navigating AI's New Frontiers in Libraries
Brian Pichman
 
Vulnerability Management: A Comprehensive Overview
Vulnerability Management: A Comprehensive OverviewVulnerability Management: A Comprehensive Overview
Vulnerability Management: A Comprehensive Overview
Steven Carlson
 

Recently uploaded (20)

Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
Mastering Board Best Practices: Essential Skills for Effective Non-profit Lea...
 
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
COVID-19 and the Level of Cloud Computing Adoption: A Study of Sri Lankan Inf...
 
Smart Mobility Market:Revolutionizing Transportation.pdf
Smart Mobility Market:Revolutionizing Transportation.pdfSmart Mobility Market:Revolutionizing Transportation.pdf
Smart Mobility Market:Revolutionizing Transportation.pdf
 
Improving Learning Content Efficiency with Reusable Learning Content
Improving Learning Content Efficiency with Reusable Learning ContentImproving Learning Content Efficiency with Reusable Learning Content
Improving Learning Content Efficiency with Reusable Learning Content
 
Intel Unveils Core Ultra 200V Lunar chip .pdf
Intel Unveils Core Ultra 200V Lunar chip .pdfIntel Unveils Core Ultra 200V Lunar chip .pdf
Intel Unveils Core Ultra 200V Lunar chip .pdf
 
kk vathada _digital transformation frameworks_2024.pdf
kk vathada _digital transformation frameworks_2024.pdfkk vathada _digital transformation frameworks_2024.pdf
kk vathada _digital transformation frameworks_2024.pdf
 
Required Documents for ISO 17021 Certification.PPT
Required Documents for ISO 17021 Certification.PPTRequired Documents for ISO 17021 Certification.PPT
Required Documents for ISO 17021 Certification.PPT
 
How UiPath Discovery Suite supports identification of Agentic Process Automat...
How UiPath Discovery Suite supports identification of Agentic Process Automat...How UiPath Discovery Suite supports identification of Agentic Process Automat...
How UiPath Discovery Suite supports identification of Agentic Process Automat...
 
Camunda Chapter NY Meetup July 2024.pptx
Camunda Chapter NY Meetup July 2024.pptxCamunda Chapter NY Meetup July 2024.pptx
Camunda Chapter NY Meetup July 2024.pptx
 
Retrieval Augmented Generation Evaluation with Ragas
Retrieval Augmented Generation Evaluation with RagasRetrieval Augmented Generation Evaluation with Ragas
Retrieval Augmented Generation Evaluation with Ragas
 
Mule Experience Hub and Release Channel with Java 17
Mule Experience Hub and Release Channel with Java 17Mule Experience Hub and Release Channel with Java 17
Mule Experience Hub and Release Channel with Java 17
 
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
UX Webinar Series: Essentials for Adopting Passkeys as the Foundation of your...
 
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
Russian Girls Call Navi Mumbai 🎈🔥9920725232 🔥💋🎈 Provide Best And Top Girl Ser...
 
Generative AI Reasoning Tech Talk - July 2024
Generative AI Reasoning Tech Talk - July 2024Generative AI Reasoning Tech Talk - July 2024
Generative AI Reasoning Tech Talk - July 2024
 
Zaitechno Handheld Raman Spectrometer.pdf
Zaitechno Handheld Raman Spectrometer.pdfZaitechno Handheld Raman Spectrometer.pdf
Zaitechno Handheld Raman Spectrometer.pdf
 
Latest Tech Trends Series 2024 By EY India
Latest Tech Trends Series 2024 By EY IndiaLatest Tech Trends Series 2024 By EY India
Latest Tech Trends Series 2024 By EY India
 
UX Webinar Series: Aligning Authentication Experiences with Business Goals
UX Webinar Series: Aligning Authentication Experiences with Business GoalsUX Webinar Series: Aligning Authentication Experiences with Business Goals
UX Webinar Series: Aligning Authentication Experiences with Business Goals
 
History and Introduction for Generative AI ( GenAI )
History and Introduction for Generative AI ( GenAI )History and Introduction for Generative AI ( GenAI )
History and Introduction for Generative AI ( GenAI )
 
Uncharted Together- Navigating AI's New Frontiers in Libraries
Uncharted Together- Navigating AI's New Frontiers in LibrariesUncharted Together- Navigating AI's New Frontiers in Libraries
Uncharted Together- Navigating AI's New Frontiers in Libraries
 
Vulnerability Management: A Comprehensive Overview
Vulnerability Management: A Comprehensive OverviewVulnerability Management: A Comprehensive Overview
Vulnerability Management: A Comprehensive Overview
 

Mariano G Tinti - Decoding SpaceX

  • 1. Decoding SpaceX Unveiling Launch Success with Data Science Mariano G. Tinti
  • 2. 2 • Executive Summary • Introduction • Methodology • Results • Conclusion • Appendix Outline
  • 3. 3 METHODOLOGIES • Collect data using SpaceX REST API and web scraping techniques​ • Wrangle data to create success/fail outcome variable​ • Explore data with data visualization techniques, considering the following factors: payload, launch site, flight number and yearly trend​ • Analyze the data with SQL, calculating the following statistics: total payload, payload range for successful launches, and total # of successful and failed outcomes​ • Explore launch site success rates and proximity to geographical markers​ • Visualize the launch sites with the most success and successful payload ranges​ • Build Models to predict landing outcomes using logistic regression, support vector machine (SVM), decision tree and K-nearest neighbor (KNN) Executive Summary
  • 4. 4 Introduction • Space Y is a new and ambitious company that aims to compete with SpaceX, the leading private space company in the world. • SpaceX was founded in 2002 by Elon Musk, with the ultimate goal of enabling human life on Mars. It has achieved many remarkable milestones in the history of space exploration. Based on public information provided by SpaceX we aim to: • Determine if a landing will be successful • Know how different variables affect the outcome of the landing, such as launch site, payload mass, booster version, etc. • Find correlations between launch sites and success rates • Describe some characteristics of the launch sites
  • 6. 6 Executive Summary • Data collection methodology: • Scraping from the Space X API and Wikipedia • Perform data wrangling • Collected data was enriched by creating a landing outcome label based on outcome data after summarizing and analyzing features. • Perform exploratory data analysis (EDA) using visualization and SQL • Perform interactive visual analytics using Folium and Plotly Dash • Perform predictive analysis using classification models • Data was standardized and divided into TEST and TRAINING sets. Four models were proposed and the most successful was selected based on the highest accuracy achieved. Methodology
  • 7. 7 • SpaceX offers a public API from where data can be obtained. • https://api.spacexdata.com/v4/rock ets/ • GitHub URL of the completed SpaceX API calls notebook: https://github.com/marianotinti/AppliedDataS cienceCapstone/blob/main/jupyter-labs- spacex-data-collection-api.ipynb Data Collection – SpaceX API Convert new dataframe to CSV Extract relevant data to dataframe Read .json response and convert into a dataframe Request API and parse the SpaceX launch data
  • 8. 8 • More data from SpaceX launches was scraped from Wikipedia GitHub URL of the completed SpaceX API calls notebook: https://github.com/marianotinti/AppliedDa taScienceCapstone/blob/main/jupyter- labs-webscraping.ipynb Data Collection - Scraping Convert new dataframe to CSV Extract relevant data to dataframe Create beautiful soup object from the response Request Falcon9 html page via HTTP GET
  • 9. 9 • Exploratory data analysis (EDA) was performed on the dataset. • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/labs-jupyter- spacex-Data%20wrangling.ipynb Data Wrangling Load the data Explore and find patterns Summarize Create the labels of mission outcomes
  • 10. 10 • To explore data, scatterplots, barplots and line charts were used to visualize the relationship between pairs of features such as: • Payload Mass X Flight Number, Launch Site X Flight Number, Launch Site X Payload Mass, orbit and Flight Number, Payload and Orbit. • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/IBM-DS0321EN- SkillsNetwork_labs_module_2_jupyter-labs-eda-dataviz.ipynb.jupyterlite.ipynb EDA with Data Visualization
  • 11. 11 • To better understand SpaceX data set, following SQL queries/operations were performed on an IBM DB2 cloud instance: 1. Display the names of the unique launch sites in the space mission 2. Display 5 records where launch sites begin with the string 'CCA’ 3. Display the total payload mass carried by boosters launched by NASA (CRS) 4. Display average payload mass carried by booster version F9 v1.1 5. List the date when the first successful landing outcome in ground pad was achieved. 6. List the names of the boosters which have success in drone ship and have payload mass greater than 4000 but less than 6000 7. List the total number of successful and failure mission outcomes 8. List the names of the booster_versions which have carried the maximum payload mass. Use a subquery 9. List the failed landing_outcomes in drone ship, their booster versions, and launch site names for in year 2015 10. Rank the count of landing outcomes (such as Failure (drone ship) or Success (ground pad) between the date 2010-06- 04 and 2017-03-20, in descending order • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/jupyter-labs-eda-sql- coursera_sqllite.ipynb EDA with SQL
  • 12. 12 We took advantage of geospatial data with the use of folium, which also allowed us to perform interactive visual analysis. • Launch sites were marked on the map with marker objects. • Added ‘folium.circle’ and ‘folium.marker’ to highlight circle area with a text label over each launch site. • Added a ‘MarkerCluster()’ to show launch success (green) and failure (red) markers for each launch site. • Calculated distances between a launch site to its proximities (e.g., coastline, railroad, highway, city) • Added ‘MousePosition() to get coordinate for a mouse position over a point on the map • Added ‘folium.Marker()’ to display distance (in KM) on the point on the map (e.g., coastline, railroad, highway, city) • Added ‘folium.Polyline()’ to draw a line between the point on the map and the launch site • Repeated steps above to add markers and draw lines between launch sites and proximities – coastline, railroad, highway, city • This way we could understand that launch sites are in close proximity to: HIGHWAYS, RAILWAYS, COASTLINES AND CITIES • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/lab_jupyter_launch_site_locati on.ipynb Build an Interactive Map with Folium
  • 13. 13 1. Added a Launch Site Drop-down Input component to the dashboard to provide an ability to filter Dashboard visual by all launch sites or a particular launch site 2. Added a Pie Chart to the Dashboard to show total success launches when ‘All Sites’ is selected and show success and failed counts when a particular site is selected 3. Added a Payload range slider to the Dashboard to easily select different payload ranges to identify visual patterns 4. Added a Scatter chart to observe how payload may be correlated with mission outcomes for selected site(s). The color-label Booster version on each scatter point provided missions outcomes with different boosters • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/Dash.py Build a Dashboard with Plotly Dash
  • 14. 14 • Four classification models were compared: logistic regression, support vector machine and decision tree. • https://github.com/marianotinti/AppliedDataScienceCapstone/blob/main/IBM-DS0321EN- SkillsNetwork_labs_module_4_SpaceX_Machine_Learning_Prediction_Part_5.jupyterlite.ipy nb Predictive Analysis (Classification) Load the dataset and create NumPy arrays Standardize data Split into TRAIN and TEST datasets Create models (Log Regression, Decision tree, K- Nearest Neighbors nad Support Vector Machine Test each model with combinations of hyperparameters Calculate Accuracy of each model and make a decision
  • 15. • Exploratory data analysis results • Interactive analytics demo in screenshots • Predictive analysis results • Accuracy for Logistics Regression method: 0.8333333333333334 • Accuracy for Support Vector Machine method: 0.8333333333333334 • Accuracy for Decision tree method: 0.8888888888888888 • Accuracy for K nearest neighbors method: 0.8333333333333334 15 Results
  • 17. 17 • CCAF5 SLC 40 is where most of recent launches were successful • Success rates (Class=1) increases as the number of flights increases Flight Number vs. Launch Site
  • 18. 18 • For launch site ‘VAFB SLC 4E’, there are no rockets launched for payload greater than 10,000 kg. • Payloads over 12,000kg seems to be possible only on CCAFS SLC 40 and KSC LC 39A launch sites. • Percentage of successful launch (Class=1) increases for launch site ‘VAFB SLC 4E’ as the payload mass increases • There is no clear correlation or pattern between launch site and payload mass Payload vs. Launch Site
  • 19. 19 • The biggest success rates happens to orbits: • ES-L1; • GEO; • HEO; • SSO. • Followed by: • VLEO (above 80%) • LFO (above 70%). GTO orbit has the lowest success rate. SO orbit has no successful launches Success Rate vs. Orbit Type
  • 20. 20 • For orbit VLEO, first successful landing (class=1) doesn’t occur until 60+ number of flights • For most orbits (LEO, ISS, PO, SSO, MEO, VLEO) successful landing rates appear to increase with flight numbers • There is no relationship between flight number and orbit for GTO Flight Number vs. Orbit Type
  • 21. 21 • Successful landing rates (Class=1) appear to increase with pay load for orbits LEO, ISS, PO, and SSO • For GEO orbit, there is not clear pattern between payload and successful launches Payload vs. Orbit Type
  • 22. 22 ✔ Success rates improved over time o Success rate (Class=1) increased by about 80% between 2013 and 2020 o Success rates remained the same between 2010 and 2013 and between 2014 and 2015 o Success rates decreased between 2017 and 2018 and between 2019 and 2020 Launch Success Yearly Trend
  • 23. 23 • Query: %sql Select distinct Launch_site from spacextbl ‘distinct’ returns only unique values • Result: All Launch Site Names
  • 24. 24 • Query: %sql select * from spacextbl where Launch_Site LIKE ‘CCA%’ limit 5; Keyword LIKE and format ‘CCA%’ returns records in the Launch Site column starting with “CCA” Limit 5 limits the number of records shown to 5 • Result: Launch Site Names Begin with 'CCA'
  • 25. 25 • Query: %sql select sum(PAYLOAD_MASS_KG_) from spacextbl where Customer = ‘NASA (CRS)’ Sum adds the results of the mentioned column and returns the total payload mass for the described customer • Result: Total Payload Mass
  • 26. 26 • Query: %sql select avg (PAYLOAD_MASS_KG_) from spacextbl where Booster_Version LIKE ‘F9 v1.1’ ‘avg’ keyword returns the average of payload mass in ‘PAYLOAD_MASS_KG’ column where booster version is ‘F9 v1.1’ • Result: Average Payload Mass by F9 v1.1
  • 27. 27 • Query: %sql SELECT min(DATE) FROM SPACEXTBL WHERE LANDING__OUTCOME='Success (ground pad)’ ‘min(Date)’ selects the first or the oldest date from the ‘Date’ column where first successful landing on group pad was achieved Where clause defines the criteria to return date for scenarios where ‘Landing_Outcome’ value is equal to ‘Success (ground pad)’ • Result: First Successful Ground Landing Date
  • 28. 28 • Query: %sql SELECT BOOSTER_VERSION FROM SPACEXTBL WHERE PAYLOAD_MASS__KG_ between 4000 and 6000 AND LANDING__OUTCOME= 'Success (drone ship)’ ; The query finds the booster version where payload mass is greater than 4000 but less than 6000 and the landing outcome is success in drone ship The ‘and’ operator in the where clause returns booster versions where both conditions in the where clause are true • Result: Successful Drone Ship Landing with Payload between 4000 and 6000
  • 29. 29 • Query: %sql SELECT COUNT(*) FROM SPACEXTBL WHERE MISSION_OUTCOME LIKE '%Success%' OR MISSION_OUTCOME LIKE '%Failure%’ The ‘group by’ keyword arranges identical data in a column in to group In this case, number of mission outcomes by types of outcomes are grouped in column ‘counts’ • Result: Total Number of Successful and Failure Mission Outcomes
  • 30. 30 • Query: %sql SELECT BOOSTER_VERSION FROM SPACEXTBL WHERE PAYLOAD_MASS__KG_ = (SELECT MAX(PAYLOAD_MASS__KG_) FROM SPACEXTBL) The sub query returns the maximum payload mass by using keywork ‘max’ on the pay load mass column The main query returns booster versions and respective payload mass where payload mass is maximum with value of 15600 • Result: Boosters Carried Maximum Payload
  • 31. 31 • Query: %sql SELECT TO_CHAR(TO_DATE(MONTH("DATE"), 'MM'), 'MONTH') AS MONTH_NAME, LANDING__OUTCOME AS LANDING__OUTCOME, BOOSTER_VERSION AS BOOSTER_VERSION, LAUNCH_SITE AS LAUNCH_SITE FROM SPACEXTBL WHERE LANDING__OUTCOME = 'Failure (drone ship)' AND "DATE" LIKE '%2015%’ The query lists landing outcome, booster version, and the launch site where landing outcome is failed in drone ship and the year is 2015 The ‘and’ operator in the where clause returns booster versions where both conditions in the where clause are true The ‘year’ keywork extracts the year from column ‘Date The results identify launch site as ‘CCAFS LC-40’ and booster version as F9 v1.1 B1012 and B1015 that had failed landing outcomes in drop ship in the year 2015 • Result: 2015 Launch Records
  • 32. 32 • Query: %sql SELECT "DATE", COUNT(LANDING__OUTCOME) as COUNT FROM SPACEXTBL WHERE "DATE" BETWEEN '2010-06-04' and '2017-03-20' AND LANDING__OUTCOME LIKE '%Success%' GROUP BY "DATE" ORDER BY COUNT(LANDING__OUTCOME) DESC The ‘group by’ key word arranges data in column ‘Landing__Outcome’ into groups The ‘between’ and ‘and’ keywords return data that is between 2010-06-04 and 2017-03-20 The ‘order by’ keyword arranges the counts column in descending order The result of the query is a ranked list of landing outcome counts per the specified date range • Result: Rank Landing Outcomes Between 2010-06-04 and 2017-03-20
  • 34. 34 Falcon 9 Launch Sites Map Fig 1 – Launch Sites Fig 3 – Launch sites located in Florida Fig 2 – VAFB SLC 4E Launch site Figure 1 on left displays the map with Falcon 9 launch sites that are located in the United States (in California and Florida). Each launch site contains a circle, label, and a popup to highlight the location, name and number of launches that happened on the launch site. It is also evident that all launch sites are near the coast. Figure 2 and Figure 3 zoom in to the launch sites to display 4 launch sites: • VAFB SLC-4E (CA) • CCAFS LC-40 (FL) • KSC LC-39A (FL) • CCAFS SLC-40 (FL)
  • 35. 35 Falcon9 Success/Failed Launch Map for all Launch Sites Fig 4 – VAFB SLC-4E Launch Site with success/failed markers Fig 3 – KSC LC -39A Launch Site with success/failed markers Fig 2 – CCAFS LC-40 Launch Site with success/failed markers Fig 1 – CCAFS SLC-40 Launch Site with success/failed markers Figure 1 Through 4 zoom in to the launch sites to display the failed/successful launches • KSC LC-39A has the greatest number of successful launches
  • 36. 36 Falcon9 – Launch Site to proximity Distance Map Fig 1 – VAFB SLC-4E Launch Site with calculated distances to: Railways, Highways and the coastline Fig 2 – VAFB SLC-4E Launch Site with calculated distances to the nearest city Figure 1 provides a zoom in view for the VAFB launch site, and shows other proximities such a coastline, railroad, and highway with respective distances from the Launch Site Figure 2 displays the proximity to the City Lompoc, which is located further away from Launch Site compared to other proximities such as coastline, railroad, highway, etc. The map also displays a marker with city distance from the Launch Site (14.09 km) In general, cities are located away from the Launch Sites to minimize impacts of any accidental impacts to the general public and infrastructure. Launch Sites are strategically located near the coastline, railroad, and highways to provide easy access to resources
  • 38. 38 Launch Success Counts For All Sites • Launch Site ‘KSC LC-39A’ has the highest launch success rate • Launch Site ‘CCAFS SLC40’ has the lowest launch success rate
  • 39. 39 • KSC LC-39A Launch Site has the highest launch success rate and count • Launch success rate is 76.9% • Launch success failure rate is 23.1% Launch Site with Highest Launch Success Ratio
  • 40. 40 • Most successful launches are in the payload range from 2000 to about 5500 • Booster version category ‘FT’ has the most successful launches • Only booster with a success launch when payload is greater than 6k is ‘B4’ Payload vs. Launch Outcome Scatter Plot for All Sites
  • 42. 42 Classification Accuracy Fig 1 – Bar plot showing the methods employe d and the accuracies achieved.
  • 43. 43 Confusion Matrix Fig 1 – Confusion matrix of the Decision tree model. Per the confusion matrix, the classifier made 18 predictions This shows that the model has a slight problem interpreting false negatives. The previous models performed worse and had equal problems with false negatives and positives. This model is more SPECIFIC towards knowing what makes the rockets NOT LAND an 88% of times
  • 44. 44 • As the numbers of flights increase, the first stage is more likely to land successfully • Success rates appear go up as Payload increases but there is no clear correlation between • Payload mass and success rates • Launch success rate increased by about 80% from 2013 to 2020 • Launch Site ‘KSC LC-39A’ has the highest launch success rate and Launch Site ‘CCAFS SLC40’ has the lowest launch success rate • Orbits ES-L1, GEO, HEO, and SSO have the highest launch success rates and orbit GTO the lowest • Launch sites are located strategically away from the cities and closer to coastline, railroads, and highways • The best performing Machine Learning Classfication Model is the Decision Tree with an accuracy of about 88%. Conclusions