Today, we will go over quickly our reading from last night.

We will then begin the process of identifying codes and categories for our analysis of tweets. Hopefully, you are already making good progress with a topic and corpus of tweets related to a hashtag. If you are still fine tuning the initial stages of collecting tweets, you may need to repeat these steps once you have your final spreadsheet in place. We will spend the first five minutes of class making sure everyone has a spreadsheet ready.

We will be undertaking four steps:

1.) Setting up your "corpus" and constraining or narrowing down the corpus of tweets. For this piece, you will create the working spreadsheet you will use for your analysis. Begin by opening the TAGS sheet with the collection of tweets that you have gathered. You want about 1,000 tweets and ideally no more than 5,000 tweets. 

Next, from the File menu, Make a Copy of the TAGS sheet, keeping the original intact. Name your copy with something you will recognize, and then set the Share settings to be public.

Next, use some filters on your spreadsheet to constrain the number of tweets. In most cases, you will likely want to filter out retweets. To do so:

• Select the column with the tweets, and then from the toolbar, select the funnel icon, and then Create New Filter View; you should now have a filter available below the name of the column.
• Click on this filter icon, and then select Filter by Condition > Text Does Not Contain; In the field that opens enter RT, and then select OK. You should now have the bulk of the retweets removed.
• Next, select and copy the entire workbook in the spreadsheet (ctrl/cmd-a), and then (ctrl/cmd-c). Then use the plus icon at the bottom left to create a new workbook. In the new workbook, paste in all of the data from your filtered view. You should now have a set of original tweets and a much smaller number of items.
• Further constrain the tweets as needed to get a number between 200-300 tweets. You might simply narrow things down by time period. Or you could take a random sample. Choose an approach, and then delete extraneous rows in the workbook to get the final set of tweets in place.
• Finally, copy the URL of the spreadsheet, and then paste it into the comment field of this posting (below) and submit the comment.

2.) Open coding to develop a list of potential categories. For this step, you will partner with a classmate. Once you have a partner, use the worksheet that I give you and follow its instructions. We will spend about 30-40 minutes on this step. Before we get started, we will take a quick look at an example.

3.) Generate categories to be used for your analysis. Use the worksheet(s) and conversation from our open coding activity to "finalize" your list of categories that you will use to code the tweets. You may be able to use some of the categories from the exercise. You may also need to add additional categories. Try to develop as sensible a list of between 5 and 7 categories.

4.) Finally, we will do additional cleanup on our corpus workbook, and then add columns for our key categories. Follow these steps:

• First, remove columns that seem to be unhelpful for your analysis; some columns have duplicate information and some will not be needed. Right/ctrl-click on a column to delete it.
• Next add columns for the final categories you have identified. Place your cursor in the column with the tweet, and then from the Insert menu, select Insert Column to the Right. Repeat this process, adding and labeling all of your categories.

Once you have completed all of these steps, you can begin the process of reading the tweets, and then entering codes to populate the categories. Remember that the process should be organic, so you may need to adjust the codes you are using in the categories as you go or even add new categories.

You can use the rest of class time to continue reading and filling in categories and codes for the tweets. 

For homework, complete the coding of all of your tweets in your final corpus before class.