SQL for Data Science Capstone Project

Justas Mundeikis

Data

This analysis is based on Olympic sports data ranging from 1896 to 2016. With over 66 different sport disciplines.

Audiance

The audience of this report are Lithuanian basketball strategist, who are interested in knowing, if Lithuanian national basketball team is becoming “older”, “lower” and “older”? As some players return every 4 years to play again for their country. They want to know, if the competitiveness of the national team might decrease due to players physique.

Hypotheses

I have raised following hypotheses:

  1. Hypothesis average height decreased among Lithuanian players at Olympics in period 1992 vs 2016
  2. Hypothesis average weight increased among Lithuanian players at Olympics in period 1992 vs 2016
  3. Hypothesis average age increased among Lithuanian players at Olympics in period 1992 vs 2016

Previous research

Exploratory data analysis can be found in my GitHub repository

H1 Height

  1. Hypothesis average height decreased among Lithuanian players at Olympics in period 1992 vs 2016

The Box plot diagram shows that the median has slightly increased over time, while the variation has decreased when comparing to 1996-2004 era.

H1 Height

H1 Height

Summary statistics table:

Summary statistics
Lithuanian Basketball players
Year avg_height sd_height median_height
1992 200.7 8.7 198.0
1996 201.7 11.1 201.0
2000 200.8 7.9 201.0
2004 201.3 8.7 199.5
2008 201.7 7.3 199.5
2012 200.3 6.4 199.5
2016 202.8 7.0 201.5

This data may suggest, that the hypothesis will be rejected.

H2 Weight

  1. Hypothesis average weight increased among Lithuanian players at Olympics in period 1992 vs 2016

The Boxplot diagram shows that the median weight has slightly increased over time, while the variation has remain approximately the same over time.

H2 Weight

H2 Weight

Summary statistics table:

Summary statistics
Lithuanian Basketball players
Year avg_weight sd_weight median_weight
1992 100.4 11.5 95.0
1996 100.6 12.9 96.0
2000 98.6 12.7 98.5
2004 100.2 9.8 97.5
2008 100.5 9.9 99.0
2012 99.4 12.2 99.0
2016 100.8 12.5 100.5

This data may suggest, that the hypothesis will be confirmed

H3 Age

  1. Hypothesis average age increased among Lithuanian players at Olympics in period 1992 vs 2016

The Box plot diagram shows that in 1922 there was the largest variation in data, then in 1996 the median age dropped and every games past that more or less the same players returned to Olympics, the median age increased. In 2016, although completely different players were playing, the median variation appears to be very similar to 1996

H3 Age

H3 Age

Summary statistics
Lithuanian Basketball players
Year avg_weight sd_weight median_weight
1992 100.4 11.5 95.0
1996 100.6 12.9 96.0
2000 98.6 12.7 98.5
2004 100.2 9.8 97.5
2008 100.5 9.9 99.0
2012 99.4 12.2 99.0
2016 100.8 12.5 100.5

Hypotheses testing

Despite exploratory data analysis showing some relative increases in basketball players physique - none of it appears to be significant

Lithuanian basketball players metrics 1992 vs 2016
student-t statiscis for height, weight, age
variable estimate1 estimate2 rel_change p.value
Age 27.33 27.75 1.52% 0.841
Height 200.67 202.75 1.04% 0.523
Weight 100.42 100.75 0.33% 0.946

Conclusions

Lithuanian basketball strategists can be calmed down, despite some players returning to play for their country, the national team does not become older, heavier and does not loose height. Thus competitiveness of national team will depend on skills and not on physique.