Story Draft

The data set that includes information on the slave sales from 1775 to 1865 contains valuable insight into the slave trade in the United States that many people may not immediately think of when they think of the history of slavery in America. It provides insights into information such as how much slaves would be sold for when they possess a particular skill set compared to slaves that possess other skills as well as no skills at all. It also allows you to compare the types of skills that female slaves possessed as opposed to the skills that were possessed by male slaves. In the visualization shown for example, you can see that female slaves that were sold possessed more domestic based skills such as “hair dresser, house servant, pastry cook, laundry, etc.”. In terms of these skills women that possessed hair dressing skills were appraised to have the highest value, being appraised at around $1,000, as compared to a female spinner that was given an average appraised value of $203. The skills that are possessed by the male slaves appear to be more skilled labor types of jobs. These jobs include skill sets such as “mechanic, brass molder, painter, cigar maker, blacksmith, construction, etc.”. By looking at the visualization we are able to see that the mechanic skill is valued the highest among the male slaves with the average slaves with skills as a mechanic were appraised to be worth around $1,258. This would be compared to a male slave possessing the skills of a pusher which is appraised at a value of $150. The visualization also allows us to compare the appraisal values of male slaves compared to the value of female slaves during the slave trade in the United States. By looking at the graph we are able to see that on average, male slaves were valued at a higher rate than female slaves. This included times when they possessed the same set of skills. For example, a male mechanic was valued at an average rate of $1,258, while a female mechanic was on average appraised at a value of $600. We are also able to see that more males possessed skills than females did. 

3 thoughts on “Story Draft

  • April 18, 2016 at 2:54 PM
    Permalink

    Re: your question in class about scaling the colors within a category, it turns out that Tableau will do that, but only with dimensions and not measures, which means that the color scaling won’t be directly by the appraised value.

    First step, differentiating colors by both gender and another category–in your case, I think it makes most sense to do it by the skills dimension, since you want each skill box to be distinguished within the gender category. If you just drop skills onto the color box, though, it’ll replace sex–so you’ll need to drop skills onto the bottom of the list in the Marks pane, and then hover over the little space to the left of it to get the drop down arrow to appear, and use that to select color.

    (Tableau is just smart enough to want assume that you only want one color scale at a time, so it buries the more advanced thing to make sure that you don’t overcomplicate things for yourself too much.)

    This will color your boxes more-or-less randomly (it scales by the alphabetical order by default, which for your purpose is pretty random). To get it to scale color by the appraised value, click the far right side of the blue Skills box in the Marks pane (you can see the little sort bars in my screenshot), and select Sort>Sort by field>Appraised Value>Average/Median. By default it’ll sort Ascending, meaning the lighter colors will be assigned to higher values and darker values to lower values, so you’ll probably want to change the sort order to Descending so that high values are dark and vice versa. You’ll still unfortunately have some randomness in the way the Female category is scaled, since Tableau will only do its sorting for all values, rather than within each category; your Male category will be correctly scaled, but your darkest boxes in the female category will be which skills have the highest appraised value overall, rather than the highest appraised value within the female category.

  • April 18, 2016 at 3:11 PM
    Permalink

    Two thoughts on your story draft: First, you might want to consider adding a filter for date so that you/your reader can see if the appraised value or # of people with that skill changes over time. You can add a filter by dragging “Date Entry” up to the Filters pane as a dimension, then selecting “quick filter”/”show filter” (depends on the version of Tableau) in the right click menu. This will give you and your reader a slider to select the year ranges you want to look at.

    Re: “more males possessed skills than females did,” right now you don’t have anything in this treemap scaled according to number of people with that skill–your boxes are sized in this version according to the average appraised value, which just tells you that more money was spent on enslaved men than enslaved women, regardless of how many people were traded. If you’d like to add number of people in to this visual while keeping the colors scaled by appraised value, use SUM(Number of Records) on your size tab, and go through the color scaling/sort by value process in my previous comment. In my attached screenshot, I have null and no talent excluded because there’s so many people in those categories they make the other skills categories hard to read, but I think you’ll get some interesting results like more male laborers being sold, but being valued less on average than the relatively small number of painters and mechanics.

    I also looked through your other sheets, and in your Appraised Value and Age by County sheets, you might want to filter out the zero values–right now they’re pulling your averages way down in each, because there are a lot of zeros for people who just had no age/value recorded.

  • April 18, 2016 at 3:19 PM
    Permalink

    Final suggestion–your Defects by State map is onto something really interesting. I’d suggest a symbol map (right next to filled map in the show me pane), with your grouped defects as color and SUM(Number of Records) as angle (how big the slice of the pie is). This will give you pie charts for each state, and could be interesting combined with a time slider filter. You’ll likely need to exclude nulls and do some grouping with your defects to narrow down how many pie slices you have to make it more readable. If your pies are quite small at first, click the size tab under Marks to make them bigger.

Comments are closed.