As it can be seen there are many hypothesis that can be made from this data. We’re going to examine factors that affects the “finishing” score of a player. As we said earlier, our data has many columns and we choose some columns among them which we thought may affect “finishing”.
Shooting skill, physics, preferred foot, heading accuracy, free kick accuracy, ball control, sprint speed, strength, positioning, penalties are some factors we thought that can be related to “finishing” level.
We examined this by pairing above attributes with “finishing” and checked the p-value for each of these pairs.
H0 = “There is no relation between the finishing score and shooting quality of a player.”
HA = “There is a relation between finishing score and shooting quality”.
Steps
First we changed the finishing column with forvet column. Details are in “Data Manipulation” part.
1) From the whole dataset, we extracted only the 2 columns; shooting and finishing.
2) Then to be able to create a crosstab, we changed the values of shooting as we did to finishing column.
3) Then, we created frequency table.
4) Next, a p value was founded which was 4.38e^-52 which is less than 0.05 so null hypothesis is rejected.
5) Finally, to support our conclusion, we also looked at the critical value. Since critical value is less than current value, again null hypothesis is rejected.
We can say that shooting quality of a player does affect the finishing score.
For example for pas accuracy, we could not reject the null hypothesis. Because our p value is 0.28 which is bigger than 0.05. Also our current value is less than critical value.
The above process is repeated for other pairs and the results are below (according to their p values):
- Strength vs Finishing : 1.15e-87
- Free kick accuracy vs Finishing: 1.7980923969199863e-25
- Ball control vs Finishing: 1.6624142938677152e-43
- Balance vs Finishing: 4.480780090446268e-21
Cea 