Step-by-step guide to TweetSets


Step 1: Select a source dataset and click Select.

Click on a dataset name to get more information.

List of source datasets with an arrow pointing to the radio button in front of a selected dataset.
Panel with arrow pointing to source dataset name and metadata displayed below.
Step 2: Limit the dataset by entering keywords, hashtags, and other criteria and then click Preview.
Dataset parameters / action tab selected, with term typed in the Contains all form field.

For stats on your dataset, see the Dataset statistics tab.

Dataset statistics tab selected with example statistics about the number of tweets and kinds of tweets in the dataset.

To view some sample tweets, see the Sample tweets tab.

Sample tweets tab selected with tweet by @Olympics.
Step 3: Click Create dataset.
Red arrow pointing to the Create dataset button.

This freezes the dataset parameters. If you did not enter any parameters, you will be taken to a page with a list of available files for download.

Step 4: Generate and download dataset exports.

Enter your email address and select the file formats you would like. Tweet JSON and Tweet CSV are only available to users on the GW network.

Form showing email address box and checkboxes for tweet JSON, tweet CSV, and tweet IDs.

You will receive an email when the dataset files have been created. Be patient. Extracts from large datasets make take hours or days to complete.

Your dataset now has a URL which you can return to later.

Arrow pointing to gray box showing message 'Link for this dataset' and a URL.

You can find your datasets under Previous datasets.

Arrow pointing to Previous datasets button on home page.

Files when ready for download.

Dataset exports tab is highlighted and two sections on the page, 'Generated tweet CSV files' and 'Generated tweet id files' contain links to .zip files.