1	41	Female	5993	8	Yes	6	Yes
2	49	Male	5130	1	No	10	No
3	37	Male	2090	6	Yes	0	Yes
4	32	Male	3068	0	No	7	No
5	30	Male	2693	1	No	1	No
6	36	Male	5237	6	No	7	No
7	29	Female	4193	0	Yes	9	No
8	31	Male	2911	1	No	5	No
9	34	Male	2661	0	No	2	No
10	28	Male	2028	5	Yes	4	Yes
11	32	Male	3298	0	Yes	6	No
12	53	Female	15427	2	No	25	No
13	38	Male	3944	5	Yes	3	No
14	36	Male	3407	7	No	5	Yes
15	34	Female	11994	0	No	12	No
16	21	Male	1232	1	No	0	No
17	34	Male	2960	2	No	4	Yes
18	44	Female	10248	3	No	22	No
19	46	Female	18947	3	No	2	No
20	44	Male	6465	2	Yes	4	No
21	30	Male	2206	1	No	10	No
22	39	Male	2086	3	No	1	Yes
23	24	Male	2293	2	Yes	2	Yes
24	43	Female	2645	1	No	5	No
25	50	Male	2683	1	Yes	3	Yes
26	35	Female	2014	1	No	2	No
27	27	Female	2341	1	No	1	No
28	30	Female	4011	1	No	12	No
29	41	Female	19545	1	No	22	Yes
30	37	Male	3022	4	No	1	No
31	35	Male	2269	1	No	1	No
32	48	Male	5381	9	Yes	1	Yes
33	28	Male	3441	1	Yes	2	Yes
34	44	Female	5454	5	Yes	4	No
35	35	Male	9884	2	Yes	4	No
36	33	Female	13458	1	Yes	15	No
37	35	Female	4014	3	Yes	2	No
38	31	Male	5915	3	No	7	No
39	32	Male	6162	1	Yes	9	No
40	38	Female	2406	1	No	10	No
41	50	Female	18740	5	Yes	27	No
42	55	Female	14756	2	Yes	5	No
43	36	Male	6499	1	No	6	No
44	36	Male	3388	0	Yes	1	Yes
45	59	Female	5473	7	No	4	No
46	29	Male	2703	0	No	5	No
47	31	Male	2501	1	No	1	No
48	32	Male	6220	1	No	10	No
49	36	Female	3038	3	No	1	No
50	35	Male	4312	0	No	15	No
51	45	Male	13245	4	Yes	0	No
52	46	Male	5021	8	Yes	4	No
53	30	Male	5126	1	Yes	10	No
54	38	Female	5329	7	Yes	13	No
55	56	Male	7260	4	No	6	No
56	23	Male	2322	3	No	0	No
57	51	Male	2075	3	No	4	No
58	30	Male	4152	1	No	11	No
59	46	Male	9619	1	No	9	Yes
60	40	Male	13503	1	No	22	No
61	51	Male	5441	0	Yes	10	No
62	30	Female	5209	1	Yes	11	No
63	46	Male	10673	2	Yes	10	No
64	32	Male	5010	1	No	11	No
65	54	Female	13549	9	No	4	No
66	24	Female	4999	0	No	3	No
67	58	Male	13872	0	No	37	No
68	44	Male	2042	4	No	3	No
69	32	Male	2956	1	No	1	No
70	20	Female	2926	1	Yes	1	Yes
71	34	Female	4809	1	No	16	No
72	37	Male	5163	5	No	1	No
73	59	Female	18844	9	No	3	No
74	50	Female	18172	3	Yes	8	No
75	25	Male	2889	1	No	2	No
76	51	Female	7484	3	No	13	No
77	24	Male	2774	0	No	5	No
78	34	Female	4505	6	No	1	No
79	34	Female	11631	2	No	11	No
80	36	Female	2835	5	No	1	No
81	30	Male	2613	1	No	10	No
82	33	Male	6146	0	No	7	No
83	56	Female	4963	9	Yes	5	Yes
84	51	Male	19537	7	No	20	No
85	31	Male	6172	4	Yes	7	Yes
86	26	Female	2368	1	No	5	No
87	58	Female	10312	1	No	40	Yes
88	19	Male	1675	1	Yes	0	Yes
89	22	Male	2523	0	No	2	No
90	49	Female	6567	1	No	15	No
91	43	Female	4739	4	No	3	No
92	26	Female	2942	1	No	8	No
93	36	Male	4941	6	No	3	No
94	51	Male	10650	2	No	4	Yes
95	39	Female	5902	4	No	15	No
96	25	Male	8639	2	No	2	No
97	30	Male	6347	0	Yes	11	No
98	32	Female	4200	7	No	5	Yes
99	45	Male	3452	5	No	6	No
100	30	Female	2632	1	No	5	No
101	32	Male	4668	0	No	8	No
102	30	Female	3204	5	No	3	No
103	30	Male	2720	0	No	5	No
104	41	Male	17181	4	No	7	No
105	41	Male	2238	2	No	5	No
106	40	Female	5605	1	No	20	No
107	35	Male	7295	1	No	10	No
108	53	Male	2306	2	Yes	7	No
109	45	Male	2348	8	No	17	No
110	32	Female	8998	1	No	9	No
111	29	Male	4319	1	No	10	No
112	58	Female	3346	4	Yes	1	No
113	40	Male	10855	7	No	12	No
114	34	Female	2231	6	No	4	No
115	22	Male	2323	1	No	2	No
116	27	Male	2024	6	No	2	No
117	28	Male	2713	1	No	5	No

Showing 1 to 13 of 1,098 entries, 7 total columns Introduction Angelica is a research statistician working for a management consulting firm in Salt Lake City. Over the years, the firm has developed a strong reputation in human resource consulting. Recently, a large medical device manufacturer, BioImplants, Inc. (or BI), has approached the company for advice on preventing employee turnover. Angelica, with quite a bit of HR consulting experience, was assigned to the project.

BI is a multinational corporation with over $1 billion in annual sales. While employee attrition is a problem in US businesses generally, it is particularly pronounced in growing and dynamic industries like healthcare. It is expensive for companies to rehire a position, costing by some estimates about 21% of the annual salary of the original position.

Angelica interviewed the Director of HR at BI, Stan Devolta, who indicated that the company has been successful in large part due to the work of the HR department and the company's commitment to its employees. It is widely recognized as a desirable place to work. However, Stan admitted that some departments within the company may be better than others at retaining employees, and that employee satisfaction is not universally high. BI would like some insight into employee turnover, and, specifically, answers to questions such as these:

- Who is leaving? - Are there changes that the company could make to retain employees?

Mr. Devolta emphasized that his ultimate goal is to get recommendations for how BI can improve attrition, along with forecasts of improvement and cost savings after implementing the recommendations. He provided Angelica with a sample of the company's HR data.

There is some time pressure. The CEO of BI wants a specific plan for employee retention within a month. Mr. Devolta has therefore stipulated a tight schedule for the analytics project: two weeks. He wants to have time to evaluate the results, revise the analytics if necessary, and fine tune the messaging.

The analytics problem for this project is supervised classification, with a target representing observed employee attrition. The company appears to have sufficient historical data to address the business problem (it has invested heavily in its data infrastructure in the last several years). However, there is missing data. Thus, one of the assumptions for the project is that the existing data contains sufficient information to model the outcome. Another assumption, of course, is that the basic conditions leading to attrition don't change in the near futurethat historical data will be representative of future data.

Based on the conversation with Mr. Devolta, Angelica suspects that the amount of overtime could be related to turnover (as suggested by the HR literature). But she makes no assumptions and wants the data to inform her of the underlying relationships.

```{r} library(tidyverse) library(caret) bi <- read_csv("bioimplants_clean.csv") bi <- bi %>% mutate(gender = as.factor(gender), over_time = as.factor(over_time), attrition = as.factor(attrition))

bi$X1 <- NULL ```

## Questions

## Q1 the average employee attrition rate at BI ```{r} attrition_rate <- mean(bi$attrition == "Yes") attrition_rate table(data$attrition) %>% prop.table() ```

## Q2 Linear probability model.

### Part a

linear probability model for attrition at the company. ```{r}

lm_model <- lm(ifelse(bi$attrition == "Yes", 1, 0) ~ ., data = bi) summary(lm_model) ```

##the effect of `num_companies_worked` on an employee's probability of leaving the company? >

### Part c. Are higher-tenured employees more likely to leave the company?

### Part d. The company is considering assigning overtime to one of its employees (employee details below). How likely will this employee leave the company with and without overtime? (With help of `coef(lm_model)%>% round(5)` command below for the values of the estimates rounded to 5th decimal.) - age: 34 years - gender: Male - monthly_income: $4,000 - num_companies_worked: 2 - years_at_company: 2 ```{r} coef(lm_model)%>% round(5)

```