-
Notifications
You must be signed in to change notification settings - Fork 0
/
MechaCarWriteUp.txt
84 lines (64 loc) · 4.18 KB
/
MechaCarWriteUp.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
MechaCar Fuel Efficiency Analysis
MechaCar provided data for a number prototype vehicles and requestd an analysis to predict
which factors would affect fuel efficiency (miles per gallon).
Overall, with the data given, the prediction of fuel efficiency is not ideal.
This conclusion was reached by first running a multilinear regression model. Assuming a probability factor of .05,
vehicle length and ground clearance provided a non-random amount of variance.
That is to say, they have a significant impact on fuel efficiency of the vehicle (miles per gallon).
Since the slope is 0.71, that is a positive slope with a fairly strong correlation
indicating that it fits the data well.
Furthermore, the y-intercept is statistically significant. This means there are other
variables and factors that contribute to the variation. Couple this with the fact that only
two variables were found to be significant contributors to fuel efficiency,
this requires asking other questions that were not asked or represented by the original dataset.
Suspension coil analysis
The design specifications for the MechaCar suspension coils dictate that the variance
of the suspension coils must not exceed 100 pounds per inch.
Does the current manufacturing data meet this design specification?
Why or why not?
The variance of the suspension coils in all lots does not exceed 100 pounds per inch.
The variance in the data was 76.23.
Mean Median Variance Std_Dev
1 1499.531 1499.747 76.23459 8.731242
However, one of the three lots was out of tolerance.
Lot 3 had a variance of 220 pounds per inch. Further questions need to be asked about
this lots manufacturing. Address the 5Ws (where, when, what was used, who, etc.) as well
as gather other facts about that lot compared to the other two.
1 Lot1 1500. 1500. 1.15 1.07
2 Lot2 1500. 1499. 10.1 3.18
3 Lot3 1499. 1498. 220. 14.8
Next, an evaluation was run to see if the suspension coil’s pound-per-inch results are
statistically different from the mean population results of 1,500 pounds per inch.
Running a t-test comparing the suspension coil data to the population mean returned:
data: Suspension_Coil_Table$PSI
t = -0.65784, df = 149, p-value = 0.5117
alternative hypothesis: true mean is not equal to 1500
95 percent confidence interval:
1498.122 1500.940
sample estimates:
mean of x
1499.531
Using a statistical significance level of 0.05, the probability-value of 0.51 is much
larger than the threshold. Therefore, the null hypothesis saying there is no statistical
difference between the sample the population is sound. The two means are statistically similar.
A future statistical analysis that could be used by MechaCar to outperform the competition would include
customer-driven factors such as:
Fuel-efficiency
Data connectivity options (Wifi, Bluetooth, satellite, etc.)
IIHS safety rating
Cost
And the always important number of cup holders
These factors are important due to the financial significance of purchasing a car. There are operating
costs like fuel efficiency to consider. Also, performance, particularly safety performance is underpinned
by numerous consumer protection acts that have established agencies charged with oversight of automobile
safety. Lastly, there has been a joke for years about cup holder ratios; prove this or disprove this once
and for all in the multilinear regression analysis.
The ultimate question to answer is, "What car features/factors sway car buyers the most?"
The null hypothesis is that the factors selected are most important to the consumer.
The alternative hypothesis is that there are factors more important to the consumer than those selected.
A multilinear regression analysis of the quantifiable (numerical) data from the new dataset would help
tease out where there are strong correlations between features and sales conversions. Collect data on
the features in car purchases, car models, as well as purchase price.
Also gathering data about the frequency of requests or actual sales with or without a specific
data connectivity technology would be useful. A chi-squared test could be used to compare the
distribution of frequencies across the two groups.