Matthew Chamberlain Project Proposal

The dataset I will be working with is “Slave Sales: 1775-1865”. The data set describes the people in slavery, giving the viewer a description of where the sale posting is from, the gender of said person, the approximate age of the slave(there are some issues with this, but I’ll describe them later), any “defects”, pertinent skill, and the price of the person.

The dataset includes information that is numeric, textual, and geographic. There are thousands of people in this list from mostly Southern States, including Georgia, Louisiana, Maryland, Mississippi, North Carolina, South Carolina, Tennessee, and Virginia. To differentiate the states, various counties are also included, and these seem to have some importance as to where slaves with certain skills are sold. In terms of numeric values, these fit into the appraised value and age columns. The minimum values for both zero, but the maximums vary greatly. While the maximum age can go up to 99 (which is probably a reporting error of some sort due to the fact that it is unlikely that someone could live to that age during this time period), the price for a person depends on skills and other factors, and the highest price for an individual thousands of dollars (again, there was one outlier, with one listing reading 525,000, which would be more than $7 million when adjusted for inflation). The range of descriptive data also varies greatly. Gender is probably the easiest to describe as it follows a binary system, male or female. Skills, however, are either left blank, or provide only a little information about the person, saying that a slave is a laborer/field-hand, or on the other side of the spectrum, a mechanic. Then there are defects, which also account for a lot of blank spaces. These descriptors can describe real ailments such as a hernia, or crippling injuries, or just state superfluous information, like whether or not the slave is a boy, or a girl, even though it’s already listed.

There are three relationships I have come up with after examining this dataset. The first is the relationship between skill set and where there was a greater concentration of slaves with skills versus those who were only laborers. For example, it seems that most slaves who were mechanics worked in Louisiana, with only a few in Mississippi and Georgia. In a similar fashion, Charleston had many slaves who were trained coopers (barrel makers). In comparison, field workers were more spread out (even though Louisiana still had a majority of them), and could be found all across the Deep South.

The second relationship comes is the type of skill a slave had versus the price for which he/she was sold. This relationship will be similar to the last one, as it pertains to skill, again. While laborers could fetch a price comparable to mechanics, the average price was far lower. There are some issues with this, as some of the information in the dataset is incomplete, or lacks a skill description. This is seen when looking for the maximum in appraised value, as after a blacksmith listed for $3500, there are no descriptions for what skills these slaves have, which could be an issue.

The third and final relationship deals with enslaved women who had specialized skill sets, and what they were sold for when compared to their male counterparts. While there are fewer women who were trained in specialized skill sets, investigating their appraised value provided interesting conclusions, as certain slaves were sold for more than males with the same skills, like hairdressing, or cooks/pastry chefs. There are also issues here, as women were not usually trained in the more intensive crafts, like smithing, barrel making, etc, which could make comparisons between the two groups difficult to make.

One thought on “Matthew Chamberlain Project Proposal

  • April 19, 2016 at 2:23 am

    I corrected some of the errors you and others have found with this data set; the next time you open your project, download the csv from course resources, open your project in tableau, and right click the name of the dataset in the upper left, then select “edit data source” and select the newly downloaded file. This will correct some of the weird outliers in the negative numbers etc that turned up.

    I mentioned this in class or office hours, but the coopers thing is really interesting, and would be worth pursing in detail–it says a lot about manufacturing and shipping, which are not usually part of the slavery narrative.

    Re: those with high value but no skills described, think about their other attributes: is there a correlation with gender and/or age? A man of a certain age might not have needed a particular skill to be perceived as valuable, for example.

    Be careful with how you word things re: gendered skills–right now you imply in the final paragraph that heavy manual labor trades like blacksmithing are more inherently valuable or skilled than things like cooking or hairdressing, which is a bit at odds with what your data is telling you how those skills were valued. Don’t project your assumptions about certain skills back onto your historical data.

Comments are closed.