Gephi
focus on data import
Clment Levallois
Gephi Support Team and Assist. Prof. @EMLYON
V 1.2 Sept 2014
1
Bio notes
Education in economics, management, history of science (Ph.D.)
Turned to digital methods for research.
data visualization, network analysis, natural language processing, web
applications and more.
Member of the Gephi Community Support team
Gephi certified trainer
https://marketplace.gephi.org/service/data-analysis/
Contact, feedback welcome: on twitter @seinecle or
www.clementlevallois.net
Clement Levallois
1. Network already formed
Clement Levallois
Network already formed?
Import is direct
Clement Levallois
Network data in Excel format
- Via the data laboratory
- Needs at least a list of
edges
- If you dont have a list of
nodes, they will be
created automatically fro
the nodes mentioned in
the edges list - just tick
the option for nodes
creation when prompted
in the edge import
wizzard!
Clement Levallois
Frequent mistakes
when importing data in the data lab:
- Take extra caution with the headers (first line) of your files! At least:
- For edges: Source,Target
- For nodes: Id,Label
- Take extra caution with the format of fields!
- Check that your numeric attributes (as opposed to textual) are not
imported as String (which means text), but as numbers indeed
(Float, Double or Integer)
- If you import a list of nodes and edges
- Import the list of nodes first!
- If you import a list of edges only
- Check that the option for the creation of missing nodes is selected
Clement Levallois
2. Data not in network format yet
Clement Levallois
The general idea
Clement Levallois
The steps
YOU START WITH
List / table of
observations
Not many solutions!
1. Extract connected
entities (a network!)
2. Format this network in a
conventional standard
3. Display the network with a
software package / in the
browser
Many solutions
Many solutions
Clement Levallois
Gaze: similarity computer
Network connecting
universities, with weak or
strong ties depending on the
similarity in terms of journals
where these univ. publish
Gaze can be downloaded from
www.clementlevallois.net
Available soon as a Gephi plugin (easier to use,
works on Macs) Contact me at
[email protected] if you need it now.
10
Clement Levallois
Creating a network from cooccurrences data in Excel / CSV
A plugin to install directly from
Gephi (Tools -> Plugins)
Network made of
authors and literary
agents
Network of article
co-authors
See next slides for further instructions on
how to use this plugin
Network formed by co-authors
and the journals where they
publish
11
Clement Levallois
1. Install the Plugin
1. Tools -> Plugin
2. Select Convert Excel and csv files in the tab Available
plugins, then install it and restart Gephi
12
Clement Levallois
2. Select Import Spigot, then Data importer
and follow the wizzard
13
Clement Levallois
Example of a dataset
This Excel file can be converted into a network where:
- Followers are connected to followers
- Followers are connected to persons being followed
- Time periods are taken into account.
Note on proper formats for the Date column:
- Can be an interval of dates, with the start date and end date separated by a comma (like above)
- Can be an interval of years, also separated by a comma (2012,2014)
- Can be a simple date (2012-12-12) or simple year (1979)
14
Clement Levallois
Create a dynamic network
via the data laboratory
Need two attributes for the nodes:
An attribute containing the apparition date for the nodes
Another one for the disappearance date for the nodes
Use the fusion function to create a time interval from these 2 attributes
(see next slide)
Same principle to create dynamic edges.
This approach has severe limits: only the topology of the network is
dynamic, not the attributes (which stay fixed). And the dynamic topology
is very limited a node can just appear and disappearm thats all.
This is a case where a programmatic approach would bring much more
flexibility.
15
Clement Levallois
The function Merge columns
Select the 2 columns which contain the
2. dates of beginning and end (date or simple
number, works as well)
1.
3.
Select Create time interval
Click on this button
16
Clement Levallois
Last step: create time interval
It is possible to leave the
end date empty.
If your temporal data are in
the format of a date, be
cautious to register the
correct format, failing to do
so will crash the creation
and probably need you to
restart Gephi to reinitialize.
Thats all! You can then switch to the Overview, activate the timeline and animate the network
17
Clement Levallois
Next step:
Gephi advanced functions
18
Clement Levallois
