last modified: 2023-04-10
1. Description of the plugin
This plugin is created by Matthieu Totet, who is a long time community member of Gephi (check his Gephi Streamer!)
You can find additional documentation on this plugin on Matthieu’s website.
This plugin enables you to:
-
Collect tweets in realtime, on the topic you choose
-
Get the connections between the users mentionned in these tweets
-
Visualize these connections in Gephi, or just export all the tweets to Excel
2. Installing the plugin
The setup of the plugin will need you to create a Twitter account, if you don’t already have one. Get your Twitter login and password ready. You also need a mobile phone number where your Twitter account can be confirmed. |
Setting up a Twitter account and creating a Twitter app is a boring task. It takes just 15 minutes, and you have to do it just once. Then you’ll have Twitter at your fingertips. Let’s go! |
3. Initializing the plugin (to be done just once)
a. Creating a Twitter account
We need now to connect to a Twitter account. If you already have a Twitter account and you are logged in, skip to here.
You have received an email by Twitter about the sign up. At some point (but now would be a good time), you must click on the link in this email to confirm it. |
The next screens of the sign up for Twitter are not important for us. Just follow the steps until you get to the page showing your Twitter profile, meaning the setup is finished.
You now have a Twitter account. We need to create a Twitter app which will authorize us to collect tweets automatically via Gephi
b. Creating a Twitter app
INFO: Almost there!!
Done!
Well done. This setup had just to be done once. Your credentials are saved and will be loaded at each run. Now we can start using the plugin.
3. Using the plugin
With the Words to follow tab, you will be able to follow one or multiple words. For hashtags , just enter the word without the hash in front of it. (e.g if you want to follow '#Gephi' just add 'Gephi'). You can enter multiple words by separating them with a comma , . E.g word1, word2 .
With the Users to follow tab, you will be able to follow the activity of one or multiple users. Any tweet from this user or retweeting or mentioning the user will be captured. You can also import all users from a
twitter list by giviing the user name
of the twitter account and the list name
to import. You can enter multiple users by separating them with a comma , . E.g totetmatt, gephi .
With the Locations to follow tab, you will be able to follow the activity of one or multiple locations. Any tweet that has been geotagged will be captured. You need to delimit the zone to track by its South West Point and the Nord East Point.
You can combine the 2 tabs, the tweet collected will be matching either the Words to follow query or the User to follow query
The Load Query File and Save Query File buttons are here to save your search queries in a file or to load it from a file. It’s convenient if you have a long list of words / users.
a. Network Logic
A Network Logic means: what should be done with an incoming tweet? How to transform it as a set of nodes and edges? It’s equivalent to map projection in the cartography world.
There are for the moment 4 Network Logics to choose from:
-
Full Twitter Network : This will represent all entities (User, Tweet, Hastags, URL, Media, Symbol etc…) as a graph.
-
User Network : This will represent the interaction between users. Any mentions, retweets or quotes between 2 users, will be represented. The size of the edge represent the number of interactions between 2 users.
-
Hashtag Network : This will create the network of Hashtags.
-
Emoji Network : Same as Hashtag Network but focused on Emoji characters ( original idea from Neel Shivdasani / @DataNeel ).
-
Bernardamus Projection (based on https://twitter.com/Bernardamus/status/1131334028043411456) : This network represents user network via hashtag present in tweets.
Note for the Emoji Network : By default, the Label
of Emoji nodes use the UTF-8 character representation. On the Overview
and the Previsualisation
panel, the lable display might not work with the default settings (you will see some square instead). To make it work, you need to use a font that support emoji. The Segoe Emoji font should be the one available by default on Win 10.
For the moment, you won’t be able to export the graph as SVG if you keep the UTF-8 character (PNG still works). But you can use the other emoji representation (html or alias) as label to export as SVG and apply post processing on it.
In the following, we use the network logic "User Network":
Be careful that if you choose very common terms, tweets will arrive fast and in large volumes.
If you don’t have enough memory (RAM) on your computer, this could make it crash.
Click on "disconnect" to stop the collection of tweets.
b. Applying a layout while the tweets arrive.
You can see the users organize spatially in real time, while the tweets are being collected.
Just run Force Atlas 2 in the "Layout" panel. This will not interrupt the collection of tweets.
Shift to the data laboratory
to view the data collected, in a spreadsheet format:
There, you can export nodes and relations ("edges") as csv files by clicking on "Export table".
c. Timeline
When you are finished with your stream, you can use the timeline feature to replay the stream of data and look at a particular time window.
This feature is still experimental and is higly subject to bug.
Extra Scripts
-
https://github.com/Minyall/gephi_twitter_media_downloader : A small script designed to take either a .csv of Tweet ids, or the export from Gephi’s TwitterStreamingImporter Plugin and download related Tweet media.
They use it !
-
Panteion University - Dept. of Communication, Media and Culture
-
Storyful
-
BBC Monitoring
-
Paris II
-
University Aberdeen
-
Athens university of economics and business / Οικονομικό Πανεπιστήμιο Αθηνών, Department of Management Science and Technology / Τμήμα Διοικητικής Επιστήμης και Τεχνολογίας
You can ping @totetmatt on twitter if you wish to expand the list ;)
to go further
Visit the Gephi group on Facebook to get help,
or visit the website for more tutorials