The website can be accessed through https://graphion.uddi.ng/. Please follow the steps under DOCUMENTATION for more information on how to locally run graphion.
Given project deadlines:
Task | Deadline date |
---|---|
Final tool/report/video | 21st of June |
Our own deadlines:
Task | Deadline date |
---|---|
Comments on video (draft) | 20th of June (before 12) |
Draft of final report | 20/21th of June |
Checking and correcting report | 21th of June |
Fix video | 21st of June |
Final touches on our tool | 21st of June |
Minimal requirements as described in VisProjectDescription.pdf and the kickoff slides.
-
At least 2 visual metaphors
- Node-link diagram with at least 3 different layouts:
- Radial
- Force-directed
- Hierarchical
- Spectral
- Orthogonal
- Circular tree
- Arc diagram (combined with matrix reorderings maybe)
- Extra: others
- Adjacency matrix with at least 5 different reordering strategies
- Clustering
- Agglomerative hierarchichal clustering
- Different ways of calculating the similarity between two clusters (aka. linkage criteria):
- Ward's method (minimizes the total within-cluster variance)
- single/minimum linkage clustering (dist(C1,C2) = min dist(Pi,Pj))
- complete/maximum linkage clustering (dist(C1, C2) = max dist(Pi, Pj))
- average linkage clustering (dist(C1, C2) = avg(dist(Pi, Pj)))
- weighted
- centroid
- median
- Different distance metrics
- Euclidean
- Cityblock
- etc..
- Agglomerative hierarchichal clustering
- Greedy algorithms
- Bipolarization
- Optimal-leaf ordering
- ...
- Clustering
- Extra: others like
- Hybrid representations
- NodeTrix
- MatLink
- 3D node-link diagram
- Hybrid representations
- Node-link diagram with at least 3 different layouts:
-
Web-based visualization tool for networks (weighted and directed graphs)
- Accessable via URL (https://graphion.uddi.ng/)
- Upload graph data (in a specified data format)
- Visual interactions from each of the 7 categories of interactions (Yi et al.)
- Data and insights should be sharable with all other people using the web tool
- 10 most recent datasets can be chosen
- Screenshots can be made of each plot
- Parameters of plots are included in screenshots
Suggested further improvements but not required:
- Multiple coordinated view design
- Different visualizations next to each other
- Brushing and linking (highlighting / selecting is synchronized between graphs)
- Edge bundling (though bold red text comes included: Only if you enjoy visualization and like to do more)
To install the development environment:
- Open the Anaconda Prompt as an administrator (solves permission issues).
- Update conda using
conda update conda
. - Clone Git repository to current location with
git clone insert_url
where insert_url is the url given by the clone button on the top right. Or download the zip file. - Go to the Git repository folder using
cd [git folder location]
. - Create a new conda environment using
conda env create -f environment.yml
.
To access the development environment:
- Open the Anaconda Prompt as an administrator (solves permission issues).
- Enter the environment through
conda activate 2ioa0
. - Make sure you go to the actual Git repository folder using
cd [git folder location]
. - Now you can work on the web app.
- After you finished a certain task, you can upload it to Github by doing the following commands in a git command prompt (or similar actions in Github desktop):
git pull origin master
to get the latest repository from Github.git add .
to add all your file changes to a commit.git commit -m "Put commit message here
to make the commit.git push origin master
to push your commit to Github.
If you decide to implement a whole new feature which requires a lot of modifications please do it on a seperate branch and create a pull request to merge it into the master branch.
To run the Flask app service locally execute the following commands (in the repository folder):
set FLASK_APP=graphion
(orexport FLASK_APP=graphion
on UNIX-like systems)flask run
The Flask service should now be running on localhost:5000.