Today I had the opportunity to give a workshop in the DigiLab in UGA’s main library. It was a packed with librarians and grad students from across campus. In just over an hour, I started with the absolute basics and showed more and more tricks that I think would help people with their reserach projects.
This was the first time I’ve ever given a presentation without powerpoint slides. As I was preparing though, it seemed silly to include detailed descriptions and screenshots when I could just switch over the Excel and show it live. I ended up putting together a handout instead, which had all of the information on it instead. The presentation (and handout) went through the following topics:
I started off by explaining what the differences were between Microsoft Office 2016, 365, and Online. Essentially, Office 2016 is stand-alone software that you buy once and keep forever but doesn’t upgrade. Office 365 is a subscription service where you have the software as long as you subscribe to it, but it upgrades all the time. Office Online is a free cloud-based version. Through UGA, we can get Office 365 for free, and a heavily reduced priced Office 2016.
I then opened up Excel and covered the absolute basics. Data entry. Moving around the spreadsheet. Cell formatting. Borders. Text formats. I’m pretty sure everyone there knew this basic stuff, but I thought I’d cover it anyway because, you never know, there might be someone who’s never seen it before.
After that, into topics that make it easier to play around with your data. Search and replace, with some extras like matching the entire cell. I also covered sorting and filtering and showed that you can filter multiple columns to get a really specific subsets of your data.
This is where I wanted to spend the most time. Pivot tables are things that a lot of people had heard of (and from the show of hands, about half the people in the room), but for some reason—and I don’t say this to be all high and mighty—I have never met anyone who knows how to use them. I learned them in my Intro to Linguistic Computing class with Monte Shelley at BYU and have used them a ton since then, but I guess people just haven’t had the chance to learn about them.
Pivot tables are dang useful. They can summarize your data in tons of fancy ways. At the very basic level, you can at least see all the unique values in a particular column in your dataset, which is good for checking typos or for copying and pasting into lookup tables (see below). But when you add columns, you can then see how many row of your data frame there are that match the row and column. In the presentation we looked at come census data and saw how many people were from each city within that county. By adding columns, we could see how many men and women there were in each city. We could then add additional columns (like fabricated favorite color data) so that we could see how many men and women liked each color within that city.
There are different ways of viewing the data as well. Instead of the raw count, we looked at how to view the proportion of men to women there were in each town. We switched to a numerical data type (fabricated weight and height data), and were able to see the average weight and height for men and women in each city, as well as the tallest, shortest, heaviest, and lightest man and women in each town. I heard some audible whoa’s from people as I showed some of this stuff, which was great to hear.
I then looked at some Excel functions. We started with some basic math, but quickly went into some functions. I showed how some can just stand on their own, like
now(). We looked at how to reference other cells in functions and how they update automatically. I showed how to create a sequential list of numbers using functions that reference each other. There are some functions like
datedif() that work on dates and others like
right() are for manipulating strings. Then there are some that make reference to ranges, like
average(). I showed how the
concat() function can be useful to string together last name and first name to create a “Last, First” column. I know there are much more complicated and useful functions than the ones I covered, but I didn’t want to intimidate anyone.
We then looked at some conditionals and how they work. As an example, I created a new column in the census data that essentially collapsed the birth state down to two: Washington or not Washington.
lookup() function is one that I use all the time, and I wanted to make sure I covered it in the workshop. When preparing for the workshop, I looked at some other site and they all mention that
lookup() is essential. What this function can do is basically link together multiple spreadsheets so it starts to act like a database. You set up a table that acts like a dictionary: alphabetic, unique values in one column, with paired information in another. You can then “lookup” some value in this table, and the function will return the information associated with it.
Why is this useful? I use it for two main purposes: a converter, and a collapser. I use it as a converter for things like turning ARPABET representations of vowels into IPA. In one column is the ARPABET pair of letters, and the other column are the IPA symbols. It’s perfect for that. I use it also to collapse data down to fewer categories. We did this in the workshop by collapsing the 50 states down to 4–5 regions. We used this lookup table to add a “region” column to the census data, and then made a pivot table with it. Pretty cool.
For the last part of this section, I showed how to handle the places where
lookup() fails, like blank cells or cells not in the dictionary. For blanks, it returns an error, and you can overcome that by wrapping the
lookup() function in
if(isblank()). But for those pesky typos,
lookup() returns the closest value, which I only discovered recently and was not happy with it. I didn’t have to demonstrate, but in the handout I show that if you do something like
=IF(COUNTIF(A1:A5,C2)>0, LOOKUP(C2, A1:A5, B1:B5), "ERROR") it’ll work great.
Next was how to do quick and dirty visualizations in Excel. I explained briefly (probably too briefly) that not all visualizations work for all kind of data, which I feel is important for people to know. I then showed how to make a bar chart, pie chart, line graph, and scatterplot. I of course used pivot tables to help summarize the data for visualization. I did say though that I don’t use Excel’s visualizations for anything because I find them ugly, not customizable enough, and not robust enough to handle what I want to do.
Bonus Tips and Tricks
In the last four minutes, I tried to cover some bonus little tips and tricks that I’ve picked up along the way. There are little things like anchoring and freezing/splitting the table. I did show conditional formatting because I use that all the time (my tables look a little pychedelic actually). In the handout I cover how to convert text-to-columns, which can be useful when importing data from somewhere else or for just splitting things up like “Last, First” into two columns. I covered paste special and how you can transpose, paste multiply, and overwriting functions.
It was a real whirlwind of a presentation but I think some people got a lot out of it. I don’t know if too many people walked away with any new skills per se, but at least people were exposed to what kinds of things Excel can do, and were given the resources (i.e. this handout) to learn how to do it themselves. I enjoyed giving the presentation, and even though I use R for most of my work nowadays, knowing the ins and outs of Excel sure is useful.
Again, you can download this handout here. Feel free to also download the three datasets I used: the vowels for one speaker and the vowels subset of larger dataset, which both come from the Linguistic Atlas of the Gulf States, and the Cowlitz County 1930 census data, which I gathered myself. Please also visit the accompanying blog post on the DigiLab website.