Exploring Astronaut Activities in SQL

About this case study

This is a notebook derived from the Datacamp Live Webinar where Justin Saddlemyer taught us the basics of regular expressions and how they can be used to categorize data among other cool things using SQL.

In this notebook, we will use string cleaning to process and analyze a database containing information on astronaut extra-vehicular activities (EVAs).

We will start by exploring the data, identifying potential issues with the columns that contain text. Next, we will take move forward to the necessary steps to clean up the data so you can draw insights the types and participants of EVAs!

Data: Extravehicular astronaut activities, including crew, durations, and purpose.

Primary Goal: Learning exploratory data analysis techniques, including data preparation and manipulation.

Secondary Goal: Get some insights into what astronauts do in space!

Why? Data will not always come pre-processed, or easy to visualize. Extracting information and restructuring it is essential to analytics.

Source of the data

Exploring our data

Let's start by looking at the table we will be working with.

SELECT * FROM evas ORDER BY duration DESC

dateobject

12/13/19720.5%

9/16/19930.5%

362 others98.9%

countryobject

USA62.9%

Russia37.1%

0

3/10/2001

USA

1

5/13/1992

USA

2

8/30/2012

USA

3

12/22/1999

USA

4

12/23/1999

USA

5

12/24/1999

USA

6

5/22/2011

USA

7

8/7/2010

USA

8

5/17/2009

USA

9

2/11/2008

USA

Let's inspect the purpose column in greater detail.

SELECT purpose FROM evas

purposeobject

Jettison suit backpacks and equip to lighten ascent1.6%

1.1%

364 others97.3%

0

First U.S. EVA. Used HHMU and took photos. Gas flow cooling of 25ft umbilical overwhelmed by vehicle ingress work and helmet fogged. Lost overglove. Jettisoned thermal gloves and helmet sun visor

1

HHMU EVA cancelled before starting by stuck on vehicle thruster that ended mission early

2

Inadequate restraints, stiff 25ft umbilical and high workloads exceeded suit vent loop cooling capacity and caused fogging. Demo called off of tethered astronaut maneuvering unit

3

Standup EVA. UV photos of stars. Ended by eye irritation that impaired vision

4

Retrieved MMOD experiment from docked Agena. Used HHMU. Lost camera and retrieved experiment. EVA ended early by unrelated spacecraft problem

5

Attached tether between Agena and Gemini. EVA ended early due to fatigue, overheating & eye sweat

6

Standup EVA. Took star photos. Agena tether ops

7

Standup EVA. Science tasks. Took star photos

8

Attached tether between Agena and Gemini. UV photos of stars. Waist tether and Dutch shoe eval

9

Standup EVA. Jettisoned equipment. Took photos

What are the most common types of EVAs?

Let's start to get a rough idea of the most popular types of EVAs astronauts take by using CASE expressions.

SELECT purpose, CASE WHEN purpose ILIKE '%photos%' THEN 1 ELSE 0 END AS photography, CASE WHEN purpose ILIKE '%collect%' THEN 1 ELSE 0 END AS collection, CASE WHEN purpose ILIKE '%construct%' OR purpose ILIKE '%install%' OR purpose ILIKE '%assembl%' THEN 1 ELSE 0 END AS installation , CASE WHEN purpose ILIKE '%replace%' OR purpose ILIKE '%fix%' OR purpose ILIKE '%repair%' OR purpose ILIKE '%servic%' THEN 1 ELSE 0 END AS repair FROM evas

purposeobject

Jettison suit backpacks and equip to lighten ascent1.6%

1.1%

364 others97.3%

photographyint64

0 - 1

60

LEASAT/SYNCOM IV repairs completed. Accidently reused depleted LiOH cans from 1st EVA.

0

61

EASE/ACCESS truss assembly and stowage demos. Deployed radar target

0

62

EASE/ACCESS truss assembly, stowage and handling demos. Simulated cable routing with rope reel and heat pipe handling with free EASE beams.

0

63

In suits in airlock at 5psi awaiting GO to manually deploy stuck solar array. Waived off when auto sys worked

0

64

Freed stuck GRO high gain antenna. Started Space Station demos (recorded crew loads, variable tension rope translation, dogbone handrails).

0

65

More Space Station demos (evals of 3 different carts and a tether shuttle, RMS translation rates and loads). Time short for all RMS/EVA tests

0

66

Capture bar/grapple fixture attach failed 3 times

0

67

Though more careful to induce less loads, capture bar/grapple fixture attach failed 5 times

0

68

3 man EVA. Manually capture/repair INTELSAT

0

69

ASEM Space Station assembly demos (module to truss mating, updated HHMU self rescue device). Aligned gimbals of Ku antenna for IV commanded locking

0

We are now ready to build this into a final query!

WITH purposes AS ( SELECT purpose, CASE WHEN purpose ILIKE '%photos%' THEN 1 ELSE 0 END AS photography, CASE WHEN purpose ILIKE '%collect%' THEN 1 ELSE 0 END AS collection, CASE WHEN purpose ILIKE '%construct%' OR purpose ILIKE '%install%' OR purpose ILIKE '%assembl%' THEN 1 ELSE 0 END AS installation , CASE WHEN purpose ILIKE '%replace%' OR purpose ILIKE '%fix%' OR purpose ILIKE '%repair%' OR purpose ILIKE '%servic%' THEN 1 ELSE 0 END AS repair FROM evas ) SELECT SUM(photography) AS count, 'photography' AS type FROM purposes UNION SELECT SUM(collection) AS count, 'collection' AS type FROM purposes UNION SELECT SUM(installation) AS count, 'installation' AS type FROM purposes UNION SELECT SUM(repair) AS count, 'repair' AS type FROM purposes ORDER BY count DESC

countint64

typeobject

0

191

installation

1

129

repair

2

16

collection

3

13

photography

How much material has been extracted from EVAs?

Skimming through the purpose column, we also saw numerous references to extracting rock/dust or geological material. In this case, it will be difficult to extract the total quantity across the columns. Regular expressions to the rescue!

We will define a pattern to extract the total pounds extracted per EVA, and then sum them up. Let's first do a sense check of the data.

SELECT purpose FROM evas WHERE purpose ILIKE '%geologic%' OR purpose ILIKE '%rock%'

purposeobject

0

First to walk on the moon. Some trouble getting out small hatch. 46.3 lb of geologic material collected. EASEP seismograph and laser reflector exp deployed. Solar wind exp deployed & retrieved. 400 ft (120m) circuit on foot. Dust issue post EVA

1

Collected 75.6 lb of geologic material. ALSEP exp deployed. 6000 ft (1800m) circuit on foot

2

Collected 94.4 lb of geologic material. ALSEP and laser reflector exp deployed. Hiked up to 0.9 miles (1.5km) from lunar module. Used MET rickshaw

3

Collected 169 lb of geologic material. ALSEP exp deployed. First use of the lunar rover. Covered 6.2 mile (10.3 km) circuit

4

Collected 208 lb of rock/dust (41lb this day). ALSEP & solar wind experiments deployed. Assembled and used lunar rover (4.2 km).

5

Collected 82 lb of rock/dust. Drove rover 11.5 km

6

Collected 90 lb of rock/dust. Drove rover 27.1 km

7

Collected 243 lb of geologic material. ALSEP exp deployed. 8 seismic explosive charges placed for later use. Used lunar rover to travel 11.4 miles (19 km) circuit with max range of 4.9 miles from lunar module.

Okay, we now know that the format of the pounds extracted is always number lbs of rock/geologic. We can construct a pattern to detect this and extract the number!

To do so, we will make use of:

•

\d+ to match one or more digits.

•

\.? to match zero or more periods.\

•

* to match zero or more digits following the optional decimal place.

•

() to specify we only want the digits.

•

\s to match the whitespace (i.e., spaces).

•

[ ] and | to specify we either want to match "geologic" or "rock".

Let's put this into action, using SUBSTRING() to extract our pattern!

SELECT purpose, SUBSTRING(purpose, '(\d+\.?\d*) lb of ((rock)|(geologic))') AS weight FROM evas

purposeobject

Jettison suit backpacks and equip to lighten ascent1.6%

1.1%

364 others97.3%

weightobject

46.30.3%

7 others1.9%

Missing97.9%

0

First U.S. EVA. Used HHMU and took photos. Gas flow cooling of 25ft umbilical overwhelmed by vehicle ingress work and helmet fogged. Lost overglove. Jettisoned thermal gloves and helmet sun visor

None

1

HHMU EVA cancelled before starting by stuck on vehicle thruster that ended mission early

None

2

Inadequate restraints, stiff 25ft umbilical and high workloads exceeded suit vent loop cooling capacity and caused fogging. Demo called off of tethered astronaut maneuvering unit

None

3

Standup EVA. UV photos of stars. Ended by eye irritation that impaired vision

None

4

Retrieved MMOD experiment from docked Agena. Used HHMU. Lost camera and retrieved experiment. EVA ended early by unrelated spacecraft problem

None

5

Attached tether between Agena and Gemini. EVA ended early due to fatigue, overheating & eye sweat

None

6

Standup EVA. Took star photos. Agena tether ops

None

7

Standup EVA. Science tasks. Took star photos

None

8

Attached tether between Agena and Gemini. UV photos of stars. Waist tether and Dutch shoe eval

None

9

Standup EVA. Jettisoned equipment. Took photos

None

Now we can use a CTE (Common Table Expression) to calculate the total amount!

WITH weights AS ( SELECT purpose, SUBSTRING(purpose, '(\d+\.?\d*) lb of ((rock)|(geologic))') AS weight FROM evas ) SELECT SUM(weight::NUMERIC) FROM weights

sumfloat64

0

1008.3

Which astronauts have the most time in EVAs?

We also have information on how much time each EVA took, as well as the astronauts who participated. Let's use this information to try and calculate totals for each astronaut!

First let's see what the maximum count of astronauts in an EVA has been by splitting the crew column. We can also use TRIM() to remove any extra whitespace from the column.

SELECT crew, SPLIT_PART(crew, ',', 1) AS first_astronaut FROM evas

crewobject

Leonid Kizim, Vladimir Solovyov2.1%

Sergei Krikalev, Anatoli Artsebarsky1.6%

202 others96.3%

first_astronautobject

Anatoly Solovyev2.7%

Jerry Ross2.4%

131 others94.9%

10

David Scott

11

Russ Schweickart

12

Neil Armstrong, Buzz Aldrin

Neil Armstrong

13

Neil Armstrong, Buzz Aldrin

Neil Armstrong

14

Allen Bean, Pete Conrad

Allen Bean

15

Allen Bean, Pete Conrad

Allen Bean

16

Allen Bean, Pete Conrad

Allen Bean

17

Ed Mitchell, Alan Shepard

Ed Mitchell

18

Ed Mitchell, Alan Shepard

Ed Mitchell

19

Ed Mitchell, Alan Shepard

Ed Mitchell

SELECT crew, SPLIT_PART(crew, ',', 3) AS third_astronaut FROM evas WHERE SPLIT_PART(crew, ',', 3) != ''

crewobject

third_astronautobject

0

Paul Weitz, Joe Kerwin, Pete Conrad

Pete Conrad

Now we know how we can extract the time for each astronaut! Let's create a CTE we can then use to piece together duration information for each astronaut.

WITH astronauts_split AS ( SELECT crew, SPLIT_PART(crew, ',', 1) AS first_astronaut, SPLIT_PART(crew, ',', 2) AS second_astronaut, SPLIT_PART(crew, ',', 3) AS third_astronaut, duration FROM evas ) SELECT * FROM astronauts_split

crewobject

Leonid Kizim, Vladimir Solovyov2.1%

Sergei Krikalev, Anatoli Artsebarsky1.6%

202 others96.3%

first_astronautobject

Anatoly Solovyev2.7%

Jerry Ross2.4%

131 others94.9%

0

Ed White

1

David Scott

2

Eugene Cernan

3

Mike Collins

4

Mike Collins

5

Richard Gordon

6

Richard Gordon

7

Buzz Aldrin

8

Buzz Aldrin

9

Buzz Aldrin

Now it's just a matter of splitting apart this table and appending each set of results to one table.

WITH astronauts_split AS ( SELECT crew, SPLIT_PART(crew, ',', 1) AS first_astronaut, SPLIT_PART(crew, ',', 2) AS second_astronaut, SPLIT_PART(crew, ',', 3) AS third_astronaut, duration FROM evas ), astronaut_durations AS ( SELECT first_astronaut AS astronaut, duration FROM astronauts_split WHERE first_astronaut != '' UNION ALL SELECT second_astronaut AS astronaut, duration FROM astronauts_split WHERE second_astronaut != '' UNION ALL SELECT third_astronaut AS astronaut, duration FROM astronauts_split WHERE third_astronaut != '' ) SELECT * FROM astronaut_durations

astronautobject

Anatoly Solovyev1.4%

Alexandr Serebrov1.4%

285 others97.3%

durationint64

0 - 536

0

Ed White

36

1

David Scott

0

2

Eugene Cernan

127

3

Mike Collins

50

4

Mike Collins

39

5

Richard Gordon

44

6

Richard Gordon

130

7

Buzz Aldrin

149

8

Buzz Aldrin

126

9

Buzz Aldrin

55

WITH astronauts_split AS ( SELECT crew, SPLIT_PART(crew, ',', 1) AS first_astronaut, SPLIT_PART(crew, ',', 2) AS second_astronaut, SPLIT_PART(crew, ',', 3) AS third_astronaut, duration FROM evas ), astronaut_durations AS ( SELECT first_astronaut AS astronaut, duration FROM astronauts_split WHERE first_astronaut != '' UNION ALL SELECT second_astronaut AS astronaut, duration FROM astronauts_split WHERE second_astronaut != '' UNION ALL SELECT third_astronaut AS astronaut, duration FROM astronauts_split WHERE third_astronaut != '' ) SELECT astronaut, SUM(duration) AS total_duration FROM astronaut_durations GROUP BY astronaut ORDER BY total_duration desc LIMIT 10

astronautobject

Jerry Ross10%

Anatoly Solovyev10%

8 others80%

total_durationint64

2311 - 3501

0

Jerry Ross

3501

1

Anatoly Solovyev

3086

2

Scott Parazynski

2825

3

Nikola Budarin

2672

4

John Grunsfeld

2527

5

Mike Lopez-Alegria

2501

6

Mike Fincke

2472

7

Dan Tani

2351

8

Victor Afanasyev

2314

9

Rick Mastracchio

2311

What is the cumulative amount of time spent in EVAs over time?

Finally, let's take a look at the cumulative time spent in EVAs by year and space program. To do so, we will need to use a window function in combination with a subquery.

SELECT TO_DATE(year::TEXT, 'YYYY') AS year, program, duration, SUM(duration) OVER(PARTITION BY program ORDER BY year) AS cumulative_duration FROM( SELECT year, program, SUM(duration) AS DURATION FROM evas GROUP BY year, program ORDER BY year, program ) AS sub ORDER BY year, program

yearobject

1965-01-012.9%

1969-01-012.9%

39 others94.1%

programobject

Space Shuttle35.3%

Mir19.1%

7 others45.6%

0

1965-01-01

Gemini

1

1965-01-01

Voskhod

2

1966-01-01

Gemini

3

1969-01-01

Apollo

4

1969-01-01

Soyuz

5

1971-01-01

Apollo

6

1972-01-01

Apollo

7

1973-01-01

Skylab

8

1973-01-01

Soyuz

9

1974-01-01

Skylab