Some data scientists show a certain snobbery about Excel files. But Excel files are still probably the most widely used format for sharing data. I think the reason is because a lot of people are used to using spreadsheets, so whether you're working in business applications or in science, people know how to use spreadsheets to collect, and put data in, and share it with each other. It's true now, even more than ever, with things like Google spreadsheets, which allow people to share them collaboratively on the internet. The problem for you is that when you're analyzing data with a scripting language like R you need to be able to extract the data out of those files so that you can do downstream analyses and processing on them. So we're gonna go back to the Baltimore camera data. I guess I must be getting a lot of speeding tickets lately cuz I've been thinking about this data set a lot. And so now we're gonna actually download the Excel version of this data set. And so in this case the URL is a little bit different. Again you just pick the URL off the website. And then what we're gonna do is we're going to actually download the file as a .xlsx file because that's the extension for Excel files. And so now we've downloaded this file. We began to use download file which remember is agnostic to the file type so it just downloads the file and puts it into this file name that you've stated here. And again we're gonna collect information on when the file was downloaded so that we can know when the analysis was performed. So the R library that is useful for this is the xlsx package. There's other packages, XLConnect that can also can be use if you would like to. They might provide a little bit more flexible interface but I like xlsx. So what we do is we actually use the command read.xlsx and we assign it again to the cameraData variable. And now we need to pass a couple of parameters so we need to pass the sheetIndex. So we need to say which sheet is the data stored on. And again we're gonna tell it that there's a header line so that the column names are labeled. And so when we load that into R again it looks almost exactly like the data that you got when you read the CSV file and you read it straight out of an Excel file. One thing to keep in mind is that you can read specific rows and specific columns. So for example, if you wanna read just the second and third columns and you wanna read just the first through fourth rows, you can actually pass those as variables to colIndex and rowIndex to this file read.xlsx. And you'll actually only read a subset of the Excel file which might be useful if you only wanna abstract a little part of the file that you have. So some further notes that might be useful for you are that you can use write.xlsx. So if you're working with people that like Excel files, you can actually write them back out after your analysis. Share them with people. It's very similar and you pass it the object that you want to write out and the file name and it will write that file out. read.xlsx2 I found is quite a bit faster than read.xlsx. But especially if you're reading subsets of rows, it might be a little bit unstable at least in my experience. XLConnect, as I mentioned, is a lot more flexible for reading, writing and manipulating Excel files. And so if you really need to do a lot of serious processing of Excel files, I've found that XLConnect might be a little bit better. If you're gonna be doing that, the XL Connect vignette is actually a really great place to start. It has lot of information about how you can do all sorts of different things. How you can manipulate and create files directly from R. In general, and for the purposes of this class but also the purposes of most analyses it's a little bit faster and easier to read files if you store them as comma separated, or tab separated flat files. They're also a little bit easier to distribute, as not everybody has Excel. It's not necessarily as cross-platform as using something like a plain text file, with just comma separated or tab separated values.
