Today I’ll show how you can visualise Twitter data (for example for a local SEO conference as I did in this example) to generate nice curated content and display trends. Hopefully you’ll find this as useful as I did.
Here’s a taster of how the graphs you create today may look:
ScraperWiki we’ll use for scraping twitter data and data hero for displaying it nicely.
Go to your new ScraperWiki account:
- Click on ‘Create a new dataset’
- ‘Search for new followers’
- Refer to this list when creating your query for the next screen, mine was: #brightonseo since:2013-09-12 until:2013-09-14
- Then the app will ask you to authorise it via your Twitter account
- Then press on ‘View as a table’
- Then ‘Download as a spreadsheet’ and download ‘all_tables.xls’
- Review the data to ensure it looks natural – ie. a big event attracts hundreds of tweets or thousands, so if you only have 50 there’s something wrong. You may have to re-run the scraper using different settings. Also remove irrelevant days (by deleting rows) if you only want data for a given day.
Go to your DataHero account, login and click on ‘Upload’ next to ‘your data’.
- Choose the .csv export from ScraperWiki
- You will be suggested charts based on it, I prefer to choose ‘Create New Chart’
- Then drag and drop data you’d like to visualise to the space on the right.
- This is the kind of a screen you should be seeing after dropping in two or even one bit of data, it really depends what you’d like to visualise, for a start I suggest just using ‘created at’ column to see at what times there were the most tweets:
- Once you’re done with a particular graph press on Export and voila you have your graphical representation of Twitter data.
That’s it – now you have your first basic graph. I encourage you to experiment with different bits of data and graphs and colours.
The data for me wasn’t 100% accurate, but it was enough to show the trends etc. Here’s the end effect of a few graphs that took me about 30 minutes or so
For full versions of the graphs I generated see the original at our Jellyfish blog.
I really hope you enjoyed this post and will give it a try. Please let me know if you figure out better ways to use the data or find any mistakes in what I’ve done – I only learned this today.