• To understand what factors affect women’s fertility and Menstrual Cycle!
• Uses Data collected from a Marquette University Study for 1665 survey responses of various factors including:
• Client’s Age, height, weight, BMI, Bleeding Intensity, Medication they take etc.

#Medication/pie chart
fertile_data <- read.csv("FedCycleData071012 (1).csv", sep = ",", header = TRUE)


#Cleaning the NA values of the raw data file

fertile_Med <-fertile_data[!is.na(fertile_data$Medvitexplain) & trimws(fertile_data$Medvitexplain) != "",]


fertile_Med <- fertile_Med%>%
  select(ClientID, Medvitexplain, Medvits, MedvitsM, everything())


#seperating each medicine in its own row, independent of client 
fertile_Med2 <- fertile_Med%>%
  separate_rows(Medvitexplain, sep = ",|and|;")%>%
  mutate(Medvitexplain = str_trim(Medvitexplain))

fertile_Med2 <- fertile_Med%>%
  mutate(category = case_when(
    str_detect(Medvitexplain,regex("vit|iron|oil|B12|B6|optivite|calcium|one-a-day|folic Acid|supplement", ignore_case = TRUE)) ~ "Supplements",
    str_detect(Medvitexplain,regex( "pain|head|migrane|migraine|aspirin|neurontin|flexeril|adderall|adderal", ignore_case=TRUE)) ~ "Pain Medicine",
    str_detect(Medvitexplain, regex( "lexapro|equate|effexor|seroquel|paxil|citalopram|wellbutrin|welbutrin|zoloft|fluoxetine|cymbalta|Cybalta|provigil|celexa|effexor", ignore_case= TRUE)) ~ "Anti-Depressents",
    str_detect(Medvitexplain, regex("thyroid|levoxyl|thyro|synthroid|syntthroid|levothroid|levothyroxine", ignore_case = TRUE)) ~ "Thyroid-related mediciation",
    str_detect(Medvitexplain, regex("insulin|biotic|reglan|antioxidant|greens|hydrochlorothiazide| alpha|minocyclen|humalog|cream|medroxyprogest|detrol|prometrium", ignore_case = TRUE)) ~ "Metabolic/Hormonal",
    TRUE ~ "Other"
  ))%>%
  select(Medvitexplain, category, everything())

fertile_check <- fertile_Med2%>%
  filter(fertile_Med2$category == "Other")

fertile_check <- fertile_check[!is.na(fertile_check$Medvitexplain) & trimws(fertile_check$Medvitexplain) != "",]

#counting number of women taking medicine 

fertile_countMed <- fertile_Med2%>%
  group_by(category)%>%
  summarise(n = n_distinct(ClientID))

Method:

Created a new column called “category” to categorize the different medicines women take. I manually searched up the ones I didn’t know and created a key using the case_when() function.

The “other” category includes special circumstance medicines relating to blood pressure, different allergy medicines, pulmonary medicine etc. There were too few of each to make into their own categories.

As expected a significantly large of women take supplements for their period/fertility at 72.5 percent. However what is even more interesting is that women are just as likely to experience a plethora of problems to take medications for such as anti-depressants, pain medicine (also expected), thyroid related medicine (a larger proportion than the pulmonary, or blood pressure related medicine in the “other” category, so I included it separately).

fertile_BMI<- fertile_data[!is.na(fertile_data$BMI) & trimws(fertile_data$BMI) != "",]

fertile_BMI <- fertile_BMI%>%
  select(ClientID, BMI, everything())

## 
## Call:
## lm(formula = MeanBleedingIntensity ~ BMI + MeanCycleLength, data = fertile_BMI)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.8552 -2.2497  0.0444  1.5190  8.2119 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)  
## (Intercept)      4.90486    2.56907   1.909   0.0592 .
## BMI             -0.02247    0.04271  -0.526   0.6000  
## MeanCycleLength  0.19046    0.07990   2.384   0.0191 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.572 on 98 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.05674,    Adjusted R-squared:  0.03749 
## F-statistic: 2.947 on 2 and 98 DF,  p-value: 0.05715

However after conducting a linear regression analysis this doesn’t hold through to a statistical significance. Although we see that there is a stronger relationship between mean cycle length and bleeding intensity with a 0.019 p value. This means that for every 1 unit increase in cycle length, bleeding intensity increases by about 0.19. However BMI doesn’t play a part in bleeding intensity in a statistically significant way. We can see this by observing the r squared (0.056), which is really low and the F statistic p value which is barely at 0.05.

## 
## Call:
## lm(formula = MeanBleedingIntensity ~ MeanCycleLength, data = fertile_BMI)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.8446 -2.2493  0.1497  1.5759  8.0492 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)  
## (Intercept)      4.36603    2.34747   1.860   0.0659 .
## MeanCycleLength  0.18931    0.07958   2.379   0.0193 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.563 on 99 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.05407,    Adjusted R-squared:  0.04452 
## F-statistic: 5.659 on 1 and 99 DF,  p-value: 0.01928

This confirms that although it is not strong there is a stronger relationship between Mean Cycle Length and bleeding intensity, although r squared is still low, meaning we can’t confirm that there necessarily a strong relationship.

Cleaned all the NA for the corresponding variables

#reading the chosen csv file dowloaded from kaggle
fertile_data <- read.csv("FedCycleData071012 (1).csv", sep = ",", header = TRUE)

#Cleaning the blank or NA cells 
fertile_clean_jes <- fertile_data[!is.na(fertile_data$Abortions) & !is.na(fertile_data$MeanBleedingIntensity), ]

#selecting the columns 
fertile_clean_jes <- fertile_clean_jes %>%
  select(Abortions, MeanBleedingIntensity, everything())


#grouping data 
fertile_group_jes <- fertile_clean_jes %>%
  group_by(ClientID) %>%
  summarise(Abortions = mean(Abortions, na.rm = TRUE), 
          MeanBleedingIntensity = mean(MeanBleedingIntensity, na.rm = TRUE)) %>%
            ungroup()

• ANOVA test applied to determine if there is a significance difference between the average mean bleeding intensity between the different abortion number levels.

##             Df Sum Sq Mean Sq F value Pr(>F)  
## Abortions    2   40.6  20.289   3.582 0.0315 *
## Residuals   99  560.8   5.665                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

• cleaned all the NA for the corresponding variables

#Cleaning the data 
fertile_clean_jes2 <- fertile_data[!is.na(fertile_data$Miscarriages) & !is.na(fertile_data$IncomeM) & !is.na(fertile_data$Weight), ] %>%
  group_by(ClientID) %>%
  summarise(Miscarriages = first(Miscarriages),
             IncomeM = first(IncomeM),
             Weight = mean(Weight, na.rm = TRUE)) %>%
  ungroup() %>%
  filter(Weight >= 80 & Weight <= 400)

• Histogram for the distribution of Miscarriage Occurrence based on Weight