For this assignment, students use Jupyter notebooks to explore, clean, manipulate, and visualize data from two separate datasets.
They are submitting their GitHub repository link from the forked repo located within the assignment instructions.
- Each Task has associated questions that they are expected to provide an answer for. Some of the questions are subjective and may require input from them on why they responded the way they did.
The assignment notebook is sectioned by lesson topic. Each topic contains a number of questions to be asked and answered of the data. Students should use the provided blank code cell as free space to explore the data and practice the methods that are required to answer the question prompts. Their responses to the questions should be inputted into the results markdown cell.
You should grade the students’ submissions not only on what values they give for their results, but also pay close attention to how they arrive at those results in their code cells. It is more essential to coach the students on improvements to how they arrive at a conclusion, than the conclusion itself.
The final set of questions asks the students to offer a solution to the business issue described in the assignment introduction. There is no single correct answer. Rather, aim to have a discussion with each student to better understand how the analysis they perform on the data sets influences their findings.
If the submission contains accurate responses to each question and adequate scratch-pad work in the code cells, the student should receive a 1 (aka, a pass). Use our solutions repository for references to what we call an accurate response. If the submission contains only responses and no code to show their work, do not give them a pass. Rather, send them a note to provide their proof of work. If the submission contains code to show their work but their responses are not all correct, let them know and offer corrective suggestions if your time permits. Below are a few contextual notes for what to pay attention to in each assignment section header.
We’ve provided some introductory EDA-type questions for the students to get started exploring this call center data. Check that your students are using the code cell blocks above the results section to ensure that they are determining their answers to these questions from using pandas methods on the dataframes. In other words, check their work to be sure they are not simply counting the number of calls or performing arithmetic to find the average.
There are no duplicate rows or extraneous columns in the datasets provided. All of the null value incoming wait time entries correspond to outgoing calls. And the “Sale” column contains an extra space trailing one of the “YES ” values. As above, pay closer attention to what the data-cleaning code looks like than the result. We want the students to answer the questions using the skills they’ve been learning in class
This section has the students compare each branch’s numbers with the combined branch numbers. We give them the code to merge both dataframes into one. These questions may be multi-step so again, keep in mind Pandas methods are being used to find the comparisons.
The key aim in this section is to have students use their best judgement to determine a chart type for each question. Their results should contain an explanation for why they chose a particular chart type. If you disagree with their reasoning, let them know.
Student responses should contain direct reference to the data to back up their conclusions.