# Analyze Mortality Patterns

We used the Trans-Atlantic Database to study mortality patterns and explored the correlation between the mortality rate and the number of enslaved people embarked, the variation of the mortality rate over time, and the distribution of the mortality rate.

### Mortality rate vs. number of enslaved people embarked

Among 34,948 total voyages, 25,184 record enslaved people disembarking in the Americas. Figure 1 below is the scatterplot of the imputed number of enslaved people embarked in Africa and died in the Middle Passage. Each circle represents a voyage.  The strong linear patterns are exciting at first sight, but we found that they are caused by imputation, which means that these linear patterns do not exist in the original database. Moreover, we found some blue circles have imputed death values smaller than zero, which is another issue of the database.

Figure 1: The number of enslaved people embarked and died among 25,184 voyages.

### Mortality rate over time

Only using the available raw data (ignoring the imputed number of enslaved people embarked and died), we visualized the mortality rate over time and focused on the period between 1800 and 1865. As we can see from Figure 2 below, there is an increase in the mortality rate after 1833.  Ironically, this increase corresponds to the Slavery Abolition Act in 1833. Possible explanations for this contradiction are the ineffectiveness of the Slavery Abolition Act and the dominance of Portuguese ships in the trans-Atlantic slave trade after the British Empire exited the market.

Figure 2: The annual mortality rate from 1800 to 1865.

### Distribution of the mortality rate

For 6,221 voyages with the number of enslaved people embarked and disembarked and landed in the Americas, we calculated the mortality rate and made a histogram. The histogram is skewed to the right with high density in the first two bars, which means the mortality rate for most voyages is below 20%. We fitted the data by a gamma distribution with the parameter shape equal to 1.03 and the parameter rate equal to 7.93 (shown in Figure 3). We can interpret the rate as the number of observations and the shape as the sum of counts from these observations, which means 1 in 8 enslaved people died during the Middle Passage.

Figure 3: Histogram of the death rate of 6,221 voyages fitted by the gamma distribution (shape = 1.03, rate = 7.93).