The Goal

So far, we have explored how to create graphs using ggplot2 and how to use tidyR to wrangle the data (which means using code to get the data ready for us to work with). However, our activities so far have been rather guided. When we get a data set at DataFest, we don’t have steps helping us decide what do to first, second, etc. All of that is up to us.

Today, you are going to be working with the energy efficiency data (the same data set we have been working with) as though we were in a competition. If you need a refresher on plotting, look back at Activity 1. If you need a refresher on data wrangling, you can look at Activity 3A.

We are also going to start our discussion of collaborative coding today. What does that mean? Well, during DataFest, everyone on your team will be coding on their own computers, but you will all be working together to complete the project. How can you do this efficiently? We will talk about some tips for this as we move through these activities. Today, we will start with practicing reproducible coding.

Setup

For this activity you will need the dplyr, ggplot2, and tidyr packages. Create a chunk and load these into R by using library(tidyverse) and library(ggplot2).

Note: This first step (loading the libraries) is something you should do before starting any analysis for this course. This is also something you should do before starting your work at DataFest.

We will continue using the energy efficiency data from last week. Load the train.csv data set into R. The easiest way to do this is just to copy the chunk that you used to load the data in Activity 2 and paste that into your Markdown file for this activity. You can also see Activity 1 Part 1 for full instructions.

The Client

At DataFest, you will be given a set of research questions from a client. These research questions will drive your work. For today, we have a client who is interested in determining what features (explanatory variables) in the data set are related to high energy use for residential buildings with at least 6000 sq feet of floor area. Recall that energy use is included in the EUI variable in the data set.

The client asks you to create visualizations and/or use a model to explore this. The goal is to identify features related to high energy use so the client can explore ways to help buildings make changes and reduce energy consumption. They ask you specifically to look at the age of the building, but want you to examine other features as well.

Data Wrangling when working in Teams

Now, data wrangling is a step that everyone on your team will need to do. If, for instance, you choose to handle missing data a certain way, or include only rows from a certain year, etc., everyone on your team needs to make the same choices before working with their data. Otherwise, you may end up fitting models to different data sets!!

Because of this, it is very important to annotate your code, which means putting short notes in your code that allow your teammates to know what steps you have taken. To do this in RMarkdown you just put a # in a chunk. Anything you put in the line with the # will be ignored when running the chunk, so you can put notes to yourself or your colleagues.

For example, in the following chunk, the only line of code the computer will run is the second one (library(ggplot2)). The first line (green, and in italics) will be treated as a comment and ignored.

# Load the ggplot2 package
library(ggplot2)

Note: The same thing works in R scripts. # marks at the start of the line tell the computer to treat the line as a comment rather than as code.

Okay, so everyone on your team has now done the same steps. Is there a way to do this if only one person is coding? Yes! One way to handle this sort of data wrangling when working with teammates is to have one person perform the wrangling and then that person send the new data set to the rest of the group, along with the code used to create it.

To take a data set from R and send it to another person, you need to be able to export a data set that you create. For us, this typically means turning a data set in R into a .csv file that you can send to a teammate. Do to this in R, we use the write.csv function.

write.csv(trainNew, "trainNew.csv", row.names = FALSE)

The first input (trainNew) is the name of the data set you want to send to your teammates. The second input (trainNew.csv) is the name of the file you will be storing the data set in. The third input (row.names = FALSE) just keeps the computer from adding in a column of 1, 2, 3, etc. to the beginning of your file.

If you get stuck with any of this, let us know!! Collaborative coding takes practice, and we’ll build up our skills gradually in this area. You will also find that different teams have different preferences when it comes to working together. This is a chance for you to figure out what works for your group.

Considering Age

Now that we have the data ready, let’s start with the variable the client asked us specifically to include - the age of the building.

Take a look at your data set (trainNew). Do we have a column that indicates building age in the data set? It turns out that no, we do not. This is common in data analysis work - sometimes the variable we need is not present, but we have the columns we need to create it in the data set. That is the case for us today.

Now that we’ve made a column for building age, we can explore this new variable.

Now, let’s take a look at what the client asked for - how does age relate to the energy use of the building.

ggplot(trainNew, aes(x = ...,
             y = ...,
             color = ...,
             shape = ...)) +
  geom_...() +
  labs(x = "...",
       y = "...",
       color = "...",
       shape = "...",
       title = "...",
       caption = "...") +
  theme_bw()

Choosing Other Variables

Now, we have a ton of variables in this data set. Our client wants to know which of them is associated with high energy usage. Does this means we have to look at all 60+ at once??? Nope. If you are familiar with variable selection techniques and want to try that, go for it. However, it is also totally reasonable to start exploring just a few variables.

As you are working in a team, it is also likely that each person will explore a feature on their own and then share their findings with the group. Let’s practice that.