In this project we will visualize the warp breaks dataset. The warp breaks data set contains weaving data from 1950 on the number of breaks in warp (a set of yarn vertically wrapped around a loom). There are measurements for looms consisting of data on three variables: the number of breaks, the type of wool, and the type of tension.

Here is a look at the data set.

If we look at the entire dataset, we see that wool and tension are string variables which contain possible values of:

• Wool: A, or B
• Tension: L, M, or H

The goal of this data science exercise is to visualize the difference in the number of breaks by loom and by groupings of wool and tension.

Step 1: Add loom numbers

As we can see from the data there is a pretty important variable missing, a loom identifier! Since it will be a useful variable to have, let us suppose the data observations are ordered such that the first observation for each group is loom 1, the second loom 2, and so on.

Let's add this loom identifier to the warp breaks dataset as a column called loom using groupby and cumcount. The values should go 1 through 9 for AL, then 1 through 9 for AM and so on. Another way to add these values is by using a for loop that repeats a count every nine steps. Feel free to use whichever way you prefer.

Show solution

Great, we did it! The first 5 rows of the warp breaks data set should look like this now:

Step 2: Combine wool and tension into one grouping variable

In this dataset there are two grouping variables: wool and tension. Since we would like to look at the distinction across both wool and tension at the same time, we will need to do one of two things, either group by both as we did in the previous step or create a new variable that uniquely identifies each group combination.

In preparation of adding colour to the figure in the next step, we will want to combine these two variables into one.

Let's use the addition operator to combine the string values in wool and tension into one variable called wool_tension.

Show solution

Our updated table with wool_tension should look like this now:

Step 3: Plot breaks for each loom

We have everything we need to visualize the data. Let's put loom on the x-axis, breaks on the y-axis, and add wool_tension as a grouping variable, we will need to include the answers from the previous two steps to update the warp breaks dataset.

Show solution

This graph should look like the following:

Okay, this graph is nice. It shows the breaks for each loom and has colour to identify groups. We see the blue colour group, AL has a higher number of breaks than the other groups. Let's aggregate the data by calculating the mean for each group and visualizing this once more with a bar graph.

Step 4: Calculate means

From this question on, the columns loom and wool_tension will be added to the dataframe for you.

In this step let's use groupby on wool_tension along with mean to calculate the mean number of breaks in each group.

Show solution

This table summary should look like this:

Step 5: Visualize with a bar graph

Now that the data is in a simpler and more compact form, let us use sns.barplot to make a bar graph with the wool_tension variable we made in step 2 on the x-axis and mean breaks we made in step 4 on the y-axis.

Show solution

Our graph should look like this:

Step 6: Improve the graph

This graph gets the message across. We definitely see that the AL group has more breaks than other groups, but this graph can be improved with a clearer delineation between groups!

Copy your code from the previous question. Make a new variable named median_breaks, where it is True if breaks is greater than the median of breaks, and False otherwise. In the plot, use this new variable as an argument for hue. Finally, update the axis and title. We have added the sns.barplot(..., dodge=False), to add the argument dodge=False to the bar plot function, since we did not teach this earlier when we learned bar plots. This is an optional parameter, but adding it will align the bars equally spaced on the x-axis and make the graph look nicer.

Show solution

Our final graph should look like this!

Very well done! You have officially completed the main content of this course! Feel free to continue to the next two lessons where we will go over what comes next in your programming journey.