Statistical Analysis of Earnings 
in terms of Days, Meal times and UBER Pro Levels


Situation: 
My earnings per delivery seems to differ as I completed more delivery runs. I tried to find whether there is a significant and meaningful difference among days, meal times and uber pro level in terms of earnings.
I'll analyse base and total earnings for each delivery run.

There are four elements in total earning of a single delivery earning. 
Base earning: This is the sole earning as you complete your delivery. Surge: Delivery supplement in rush hours. Boost: Delivery supplement in idle hours. Tip: Additional complementary payment by customer. We will analyse two different earning types: Base earning and Delivery earning(base+surge)


1. Base Earning analysis


1.1 T-Test of Base Earning for Weekdays and Weekend's Earnings


In this test, I tried to find whether there is a significant difference between Weekdays and Weekends in terms of delivery base earnings.
Weekends includes Friday, Saturday and Sunday, and weekdays are Monday to Thursday. Earnings represents base earning for each delivery run.


Ho: There is no significant diffence between weekday and weekend base earnings.
Ha: There is a significant diffence between weekday and weekend base earnings.

1. First, we analyse the sample sizes and means. There is slight difference between weekday and weekend earning's means. To see if this difference is significant, we apply independent t test.


Group StatisticsGroup Statistics, table, 1 levels of column headers and 2 levels of row headers, table with 6 columns and 4 rows
  day_type N Mean Std. Deviation Std. Error Mean
delivery_base_earning Weekday 527 11.1678 5.30852 .23124
Weekend 706 11.7971 5.50313 .20711
Independent Samples TestIndependent Samples Test, table, 3 levels of column headers and 2 levels of row headers, table with 11 columns and 6 rows
  Levene's Test for Equality of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference
Lower Upper
delivery_base_earning Equal variances assumed .429 .513 -2.017 1231 .044 -.62928 .31206 -1.24151 -.01705
Equal variances not assumed     -2.027 1154.224 .043 -.62928 .31043 -1.23836 -.02020

2.In t-test we need to see if variances of two groups are comparable. As we check the Levine's test result, p value is greater than our confidence level(95%), F-Sig. 0.513 > 0.05.
Therefore there is no difference between variances, so that we use the top line for analysis.

3.In this phase we will look at p value (sid.2 tailed) to interpret our result. Since p value(0.044) < 0.05, we may conlude our analysis as "The difference of means is meaningful".


Conclusion: We can say weekend base earnings are slightly higher than weekday earnings.


1.2 T-Test of Base Earning for Lunch and Dinner Time


In this test, I tried to find whether there is a significant difference between lucnh and dinner time in terms of delivery base earnings.
Lunch time is before 17:00 and dinner time is after.


Ho: There is no significant diffence between lunch and dinner time base earnings.
Ha: There is a significant diffence  between lunch and dinner time base earnings.

1. There is slight difference between lunch and dinner time base earnings.


Group StatisticsGroup Statistics, table, 1 levels of column headers and 2 levels of row headers, table with 6 columns and 4 rows
  period N Mean Std. Deviation Std. Error Mean
delivery_base_earning Lunch 377 11.2841 5.59762 .28829
Dinner 856 11.6356 5.35076 .18289
Independent Samples TestIndependent Samples Test, table, 3 levels of column headers and 2 levels of row headers, table with 11 columns and 6 rows
  Levene's Test for Equality of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference
Lower Upper
delivery_base_earning Equal variances assumed 1.364 .243 -1.048 1231 .295 -.35150 .33548 -1.00967 .30667
Equal variances not assumed     -1.030 690.354 .304 -.35150 .34141 -1.02182 .31883

2. As we check the Levine's test result, p value is greater than our confidence level(95%), F-Sig. 0.0.243 > 0.05.
There is no difference between variances, so that we can progress with top line.

3.In this phase we will look at p value (sid.2 tailed) to interpret our result. Since p value(0.295) > 0.05, We accept the Ho and say there is no significant difference.


Conclusion: We can say there is no significant difference between lunch and dinner time base earnings.


1.3 ANOVA of Base Earning for UBER pro level


In UBER Eats, as you complete single delivery run, you earn some exrerience points. Regarding your points your UBER pro level upgrades and provide you some additional amenities, such as fuel discount.There are 4 pro levels degrees of which are green, gold, platinum and diamond respectively. With this analysis, it can be deduce if there is significant diffence in base earnings.


Ho: There is no significant diffence between uber pro levels in terms of base earnings.
Ha: There is a significant diffence between uber pro levels in terms of base earnings.


Test of Homogeneity of VariancesTest of Homogeneity of Variances, table, delivery_base_earning, 1 layers, 1 levels of column headers and 0 levels of row headers, table with 4 columns and 4 rows
delivery_base_earning
Levene Statistic df1 df2 Sig.
2.182 3 1229 .088

1. Test of homogeneity shows us Sig. level (.088) is over 0.05, which means there is no difference between the variences of four groups and they can be comparable.


ANOVAANOVA, table, delivery_base_earning, 1 layers, 1 levels of column headers and 1 levels of row headers, table with 6 columns and 6 rows
delivery_base_earning
  Sum of Squares df Mean Square F Sig.
Between Groups 378.870 3 126.290 4.322 .005
Within Groups 35914.024 1229 29.222    
Total 36292.895 1232      

2.P value 0.005< CI level (0.05), indicates that there is a meaningful difference between group's means. To analyse this difference we'll examine the post hoc tests.



Post Hoc Tests


3. In descriptive analysis, results there is a slightly incresing difference in means as the pro level increases. Multiple Comparision table reveals the significance of these differences.


DescriptivesDescriptives, table, delivery_base_earning, 1 layers, 2 levels of column headers and 1 levels of row headers, table with 9 columns and 9 rows
delivery_base_earning
  N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum
Lower Bound Upper Bound
Green 130 10.2682 4.50794 .39537 9.4859 11.0504 4.62 26.58
Gold 197 11.1620 5.60044 .39901 10.3751 11.9489 5.00 36.87
Platinum 237 11.2750 5.52286 .35875 10.5683 11.9818 5.00 36.26
Diamond 669 11.9705 5.46444 .21127 11.5556 12.3853 5.00 34.55
Total 1233 11.5281 5.42757 .15457 11.2249 11.8314 4.62 36.87
Multiple ComparisonsMultiple Comparisons, table, Dependent Variable, delivery_base_earning, Scheffe, 1 layers, 2 levels of column headers and 2 levels of row headers, table with 7 columns and 18 rows
delivery_base_earning
Scheffe
(I) uber_pro_level (J) uber_pro_level Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval
Lower Bound Upper Bound
Green Gold -.89383 .61084 .544 -2.6038 .8161
Platinum -1.00687 .58999 .406 -2.6585 .6447
Diamond -1.70231 .51814 .013 -3.1528 -.2519
Gold Green .89383 .61084 .544 -.8161 2.6038
Platinum -.11304 .52119 .997 -1.5720 1.3460
Diamond -.80848 .43820 .334 -2.0352 .4182
Platinum Green 1.00687 .58999 .406 -.6447 2.6585
Gold .11304 .52119 .997 -1.3460 1.5720
Diamond -.69544 .40863 .408 -1.8394 .4485
Diamond Green 1.70231 .51814 .013 .2519 3.1528
Gold .80848 .43820 .334 -.4182 2.0352
Platinum .69544 .40863 .408 -.4485 1.8394

4. As we examine the significance values in multiple comparison table, we can say only meaningful difference is between green and diamond pro levels, which is $1.7. Other groups have no meaningful difference.


Conclusion: Green level, when I made my first deliveries, my base earnings were slightly low. As I graded up in pro levels, base earnings per order didn't affected by pro levels.


2. Delivery Earning analysis (Base+Surge)


2.1 T-Test for delivery earning  for Lunch and Dinner Time


In this test, I tried to find whether there is a significant difference between lucnh and dinner time in terms of delivery earnings(base+surge).
Lunch time is before 17:00 and dinner time is after.


Ho: There is no significant diffence between lunch and dinner time earnings.
Ha: There is a significant diffence  between lunch and dinner time earnings.

1. There is slight difference between lunch and dinner time delivery earnings as we look at the means.


Group StatisticsGroup Statistics, table, 1 levels of column headers and 2 levels of row headers, table with 6 columns and 4 rows
  period N Mean Std. Deviation Std. Error Mean
delivery_earning Lunch 377 11.2841 5.59762 .28829
Dinner 856 11.9868 5.68636 .19436
Independent Samples TestIndependent Samples Test, table, 3 levels of column headers and 2 levels of row headers, table with 11 columns and 6 rows
  Levene's Test for Equality of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference
Lower Upper
delivery_earning Equal variances assumed .045 .831 -2.009 1231 .045 -.70271 .34982 -1.38902 -.01640
Equal variances not assumed     -2.021 729.208 .044 -.70271 .34769 -1.38530 -.02012

2. As we check the Levine's test result, p value is greater than our confidence level(95%), F-Sig. 0.831 > 0.05.
There is no difference between variances, so that we can progress with top line.

3.In this phase we will look at p value (sid.2 tailed) to interpret our result. Since p value(0.045) < 0.05, We reject the Ho and accept Ha, that is there is significant difference between lunc and dinner time earnings.


Conclusion: It can be deduduced dinner time delivery earnings are slightly over lunch time earnings.


2.2 T-Test of Delivery  Earning for Weekdays and Weekend


In this test, I tried to find whether there is a significant difference between Weekdays and Weekends in terms of delivery earnings (base+surge).
Weekends includes Friday, Saturday and Sunday, and weekdays are Monday to Thursday. Earnings represents base earning for each delivery run.


Ho: There is no significant diffence between weekday and weekend earnings.
Ha: There is a significant diffence between weekday and weekend earnings.

1. First, we analyse the sample sizes and means. There is a difference between weekday and weekend earning's means. To see if this difference is significant, we apply independent t test.


Group StatisticsGroup Statistics, table, 1 levels of column headers and 2 levels of row headers, table with 6 columns and 4 rows
  day_type N Mean Std. Deviation Std. Error Mean
delivery_earning Weekday 527 11.3187 5.44251 .23708
Weekend 706 12.1103 5.80866 .21861
Independent Samples TestIndependent Samples Test, table, 3 levels of column headers and 2 levels of row headers, table with 11 columns and 6 rows
  Levene's Test for Equality of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference
Lower Upper
delivery_earning Equal variances assumed 1.891 .169 -2.432 1231 .015 -.79165 .32555 -1.43034 -.15296
Equal variances not assumed     -2.455 1169.783 .014 -.79165 .32249 -1.42437 -.15894

2.In t-test we need to see if variances of two groups are comparable. As we check the Levine's test result, p value is greater than our confidence level(95%), F-Sig. 0.169 > 0.05.
Therefore there is no difference between variances, so that we use the top line for analysis.

3.In this phase we will look at p value (sid.2 tailed) to interpret our result. Since p value(0.015) < 0.05, we may conlude our analysis as "The difference of means is meaningful".


Conclusion: We can say weekend earnings are slightly higher than weekday earnings.


2.3 ANOVA of Delivery Earning for UBER pro level


I will examine pro levels in terms of delivery earning.


Ho: There is no significant diffence between uber pro levels in terms of delivery earnings.
Ha: There is a significant diffence between uber pro levels in terms of delivery earnings.


Test of Homogeneity of VariancesTest of Homogeneity of Variances, table, delivery_earning, 1 layers, 1 levels of column headers and 0 levels of row headers, table with 4 columns and 4 rows
delivery_earning
Levene Statistic df1 df2 Sig.
2.675 3 1229 .046

1. Test of homogeneity shows us Sig. level (.046) is less than 0.05, but we can round up it to 0.5 and continue as we assume variances  are similar.


ANOVAANOVA, table, delivery_earning, 1 layers, 1 levels of column headers and 1 levels of row headers, table with 6 columns and 6 rows
delivery_earning
  Sum of Squares df Mean Square F Sig.
Between Groups 380.597 3 126.866 3.980 .008
Within Groups 39176.187 1229 31.876    
Total 39556.784 1232      

2.P value 0.008< CI level (0.05), indicates that there is a meaningful difference between group's means. To analyse this difference we'll examine the post hoc tests.



Post Hoc Tests


3. In descriptive analysis, results there is an incresing difference in means as the pro level increases. Multiple Comparision table reveals the significance of these differences.


DescriptivesDescriptives, table, delivery_earning, 1 layers, 2 levels of column headers and 1 levels of row headers, table with 9 columns and 9 rows
delivery_earning
  N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum
Lower Bound Upper Bound
Green 130 10.4115 4.71511 .41354 9.5933 11.2297 4.62 26.58
Gold 197 11.3866 5.79557 .41292 10.5723 12.2009 5.00 36.87
Platinum 237 11.6970 6.02074 .39109 10.9265 12.4674 5.00 39.76
Diamond 669 12.1764 5.62953 .21765 11.7490 12.6037 5.00 34.55
Total 1233 11.7720 5.66637 .16137 11.4554 12.0886 4.62 39.76
Multiple ComparisonsMultiple Comparisons, table, Dependent Variable, delivery_earning, Scheffe, 1 layers, 2 levels of column headers and 2 levels of row headers, table with 7 columns and 18 rows
delivery_earning
Scheffe
(I) uber_pro_level (J) uber_pro_level Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval
Lower Bound Upper Bound
Green Gold -.97506 .63798 .506 -2.7610 .8109
Platinum -1.28542 .61620 .226 -3.0104 .4395
Diamond -1.76483 .54116 .014 -3.2797 -.2499
Gold Green .97506 .63798 .506 -.8109 2.7610
Platinum -.31036 .54434 .955 -1.8342 1.2135
Diamond -.78977 .45767 .395 -2.0709 .4914
Platinum Green 1.28542 .61620 .226 -.4395 3.0104
Gold .31036 .54434 .955 -1.2135 1.8342
Diamond -.47941 .42679 .738 -1.6741 .7153
Diamond Green 1.76483 .54116 .014 .2499 3.2797
Gold .78977 .45767 .395 -.4914 2.0709
Platinum .47941 .42679 .738 -.7153 1.6741

4. As we examine the significance values in multiple comparison table, we can say only meaningful difference is again between green and diamond pro levels, which is $1.7. Other groups have no meaningful difference.


Conclusion: Similarly, except from green level, there is no difference in other pro levels, which means pro level doesn't affect the delivery earning. 


3.Conclusion:


1. Both base earnings and delivery earnings are slightly higher in weekends.
2. For base earnings lunch time and dinner time earns similar, but with the surge effect dinner earnings gets higher.
3. Pro level doesn't affect on earnings. The more I get experienced doesn't mean the more I  earn per delivery.