Photo of CCICADA

Michael N. Tsamis


Office: CoRE Room 450
Email: Photo of MTsamis

I obtained a B.S. in Computer Information Systems with an applied specialization in Criminal Justice and minor in English from John Jay College of Criminal Justice in June 2011. The courses I have taken have allowed me to learn the fundamentals of how to secure computers and their data. Based on my academic performance, I was accepted into my college's DHS Undergraduate Career Development Program funded by an HS-STEM Career Development Grant. This program provided me the opportunity to work in fields related to homeland security and conduct research related to anomaly detection and cyber security under the guidance of both Dr. James Abello and Dr. Bilal Khan who are active in this research area. Upon concluding my summer internship with DIMACS/ CCICADA, I hope to have a career related to cyber security or computer forensics.

Current Project:

Anomaly Detection

This project addresses the problem of finding persistent patterns in evolving networks. A central question is the characterization of patterns that can be used as the basis to detect anomalous activities in time-evolving networks. One of the datasets we will be working with is data from Twitter.
Mentor: Dr. James Abello
A joint project between myself and Maria Taranov


Week 1:

I began this week by familiarizing myself with ASK-GraphView and Cytoscape which are two programs used for graph visualization. I also reviewed research papers on graph visualization software, matrices, and tensors. I met with Dr. Abello and discussed our next step which involves creating a client-server architecture for a user (the client) to have the ability to use ASK-GraphView or Cytoscape to select any node in a given graph and assign it to a new cluster. The necessary changes would then be sent to the server where a graph algorithm will be applied to calculate any new edges or vertices. The server would then create a new file with the node changes and send it back to the client.

Week 2:

I created two very basic client-server programs using C which simply allows the client to send a message to the server while the server responds to every message it receives. By the end of this week, I will be modifying these programs so that they are capable of sending and receiving actual files that contain graph information. In addition, Dr. Abello informed me that we will be detecting anomalies in data from Twitter. I am currently working on finding a good API capable of obtaining Twitter data. In the meantime, a graduate student at DIMACS provided me with a sample of Twitter data already graphed in ASK-GraphView so that I can begin looking for patterns in "retweets" and "Twitter hashtags".

Week 3:

I found an API capable of retrieving Twitter data from users whos tweet contains a specific search term. The following data is retrieved:

  • The user who tweeted the message
  • The Tweet message
  • A URL to the user's profile picture
  • A timestamp for the tweet
  • The user's ID number
  • The language the tweet is written in
  • The user IDs for any users linked within the Tweet via the "@" symbol
  • A URL to the Tweet
After configuring the API, I began collecting Twitter data from Tweets which contained "#Libya", "#Egypt", "#Yemen", "#Tunisia", and "#Syria" which are all countries that are currently undergoing revolutions between civilians and their government and may concern homeland security. I have successfully collected over 150,000 pieces of Twitter data which contains the search terms named above. I was also able to convert the data into a format that can be read by my C++ program which will convert the data into an undirected graph.

Week 4:

I was able to get my C++ program to read the Twitter data and format it into an undirected graph in which the username is the source node, the Tweet message is the destination node, and the timestamp is the edge/ interaction. In addition, Maria and I obtained a Java program which is able to read the undirected graph outputted by my C++ program and place the edges and vertices into clusters using network modularity function. In the meantime, Maria is working on a Java program which will calculate the membership ratio for each of the clusters. This is all being done so that we can effectively find patterns and anomalies in Twitter networks.

Week 5:

Coming Soon...

  • J. Abello, F. van Ham and N. Krishnan, ASK- GraphView: A Large Scale Graph Visualization System, IEEE Transactions on Visualization and Computer Graphics, Vol. 12, No. 5, September/October 2006.
Previous Work:

Sonification of Network Traffic


The ability to monitor normal and abnormal network traffic in real time is vital since a network attack can occur and be completed in seconds. One novel way to render network traffic is through its acoustic representation. The Centaur project (Centralized Auralization) explores dynamic sonic interpretation of live TCP/IP network packets as musical notes of varying pitch, volume, duration and instrumentation. Centaur renders a stream of TCP packets in a manner that reflects the system's continuously updated beliefs concerning the network state, e.g. whether a “port scan” or “distributed denial of service attack” (DDoS) is taking place. Additionally, the system is capable of mapping normal web traffic contents into musical notes. Multiple Centaur sensors can be deployed to monitor machines across the wide area Internet; a central Centaur server receives musical note commands over the network from deployed sensors. The advantage of this architecture is that the guarded machines need not cohabit the same local area network (LAN). Future work will evaluate the extent to which such acoustic renderings enable administrators to more effectively (and viscerally) sense shifts in patterns of network utilization.
Mentor: Dr. Bilal Khan

  • This research project will be presented at the American Society of Criminology Conference on November 16-19, 2011 in Washington, D.C.

Constellation System


This project involves the design and development of a multi-user multi-role web-based system that will be utilized by John Jay College's PRISM Program. PRISM offers mentoring and stipends to undergraduate students with research interests relating to science. The project's aim was to enable students to search and apply online for PRISM-related opportunities instead of requiring paperwork. Once a student uploads their application, references can submit their recommendation letters electronically. In addition, mentors have access to the website to approve students/projects they wish to supervise. Once all of the components of the student's application are completed, the reviewers can evaluate the full application and submit their decision. This project makes use of a MySQL database which stores user information, application data, and active solicitations that PRISM chooses to make available to students. The website itself, designed with HTML, PHP, Javascript, and CSS, allows users to upload, download, view, and submit data to PRISM
Mentor: Dr. Bilal Khan
A joint project between myself and Mateusz Opalinski