Gabriel Rizk Journal

Total Hours: 151.5 Hours

Week 1: 10.5 Hours

1/24: Read about medical uses of VR. Most interesting aspect was the use of VR for pain mitigation during complex procedures.

https://www.nbcnews.com/mach/science/3-ways-virtual-reality-transforming-medical-care-ncna794871 (2.5 hours)

1/26: Read an extremely interesting paper regarding specific uses cases of medical VR.

https://journals.lww.com/anesthesia-analgesia/fulltext/2007/12000/The_Analgesic_Effects_of_Opioids_and_Immersive.43.aspx (1.5 hours)

Copy and paste the link (clicking isnt working.)

1/27: Setup Steam on my computer and read more into last years wiki (2 hours)

1/28: Went to the Art of Data: Visualization talk in Boston and had dinner with one of the organizers  (4 hours)

Gained alot of insight into common data visualization techniques used in the industry

Week 2: 7 Hours

1/31: Super interesting articleI I read today

https://vrtodaymagazine.com/medical-virtual-reality/ 

 Played with VR and went to the YURT. (3 hours)

Homework for 1/31: Potential Project Names:

• MRI Data Visualization

• Quiqup London Delivery Clusters

• Project F.L.Y

2/4: Played around with Google Earth VR to research how I could implement another potential project idea (2 hours)

Homework for 2/5: 

Project Ideas and Things to Do

• MRI Data Visualization

  • Take work I've done with Professor Badre in the CLPS department to visualize the MRI data that I've completed analysis on.

  • Cluster nuerons where potentially memories are stored based on their triggers from an experiment in the lab.

  • Create a tutorial for others to follow to do the same!

  • For the in class activity, I can display the MRI data and the see if the class can pair clusters to specific memories (after I explain the expirement)

    • Mostly a guessing game, but would be a fun activity to partake in.

• Quiqup London Delivery Clusters

  • Make an overlay of a map of London to display order clusters and their corresponding delivery methods

    • Would help show why and where smaller modes of transportations are used (narrow streets, close to point of pickup, etc)

  • Create a tutorial for others to do the same. Would be interesting as data clustered in locations could look like lights in the night sky (look cool), and give actionable insight for organizations to use

  • I would need  to recluster the data and do some more analysis of it.

  • For the in class portion, I would create a game where we people would try to find out reasons why the data clusters as it could using the map, and see if the insight gained is comparable to what I found from the NLP tool I built.

• Project F.L.Y

  • Visualize a drone and its movements purely from its data 

    • I took Intro to Robotics last semester and built a drone. I wanted to see if I could visualize a drone's movements purely from its data

  • I would like to create a tutorial for this, for people to be able to use their own drone data to simulate the movements.

  • I would need to create use the algorithms we learned in the class to create this product. It could be we see things from the perspective of the drone in VR

Week 3: 8 Hours

2/5 Researched how to visualize MRI data using Paraview (1 hour)

Read an interesting paper regarding its applications: exploring as a potential project

Homework for 2/7: 

Project Draft Plan (May change project)

Title: Visualize MRI Data

Description: I worked with professor Badre in the CLPS department to determine if certain images are located in specific spots in the brain.

Pre-activites: Research how to model MRI data, finish data analysis

Goal-Objective: Make the data easier to understand by clustering the points in a 3d space overlayed over a human brain.

Contribution: Find a modeling software to use. (2.5 hours)

2/10: Explore another research idea, consulted my former employer for use of their data for another idea shown below (2 hours)

Homework for 2/12: 

Pre-project Milestones:

• Make a list of the objectives and how to complete them

• Make a list of the softwares I'm going to use and in what phases

• Redo the clustering

2/12:

FIRST PROJECT DESCRIPTION:

Title - Quiqup London Delivery Clusters

Description: This past summer, I interned at a company called Quiqup in London. Quiqup is a live courier service, where anyone can get a courier to essentially do anything through an app. If you want your laundry picked up, they can get it to your house in an hour. If you want groceries, same thing. They get everything from takeout to TVs. I worked as a data scientist, and I built a tool to assign orders to different modes of transportation using a Neural Network. The problem required the complexity of a nueral network because most of their orders were text. Someone would enter the spot and say "Pick up the laundry from this location. I have these items," and they sent their receipt. The computer doesn't know how big a shirt or a suit is, and if there are several of them, they won't fit with a scooter or a bike (most common MOTs). This required a lot of NLP, but also involved clustering because locations had common MOTs. I want to display these clusters on an overlay of Google Earth with a map of London to better understand why these clusters occurred (grocery stores, narrow streets, etc)

Contributions: my main contribution is to create a VR map of London with Google Earth and overlay it with the clusters, so that I can visualize where and why these clusters occur. I may intern there again this summer, so this tool could be used to better optimize MOTs and provide better rulesets for when the tool isnt able to make a hard cut decision if things are borderline between two MOTs.

Prior Knowledge: I have alot of experience with data analysis and Neural Networks. The math and clustering is mostly done, now I just have to visualize it in VR

Milestones:

• Week 1 

  • Find software that allows me to visualize clusters of data 

• Week 2 

  • Learn how to work with Open Street Map

• Week 3

  • Conduct the data clustering again as it wasn't perfect the first time I used, and was not the main feature of the project (Nueral Network was)

• Week 4

  • Visualize the data clusers using a different software on my computer (2D for now)

• Week 5

  • Begin creating the tutorial on how to visualize the data

• Week 6

  • Work with Google Earth VR to begin displaying the data clusters in specific locations.

Deliverable: My project will have two concrete deliverables. First, I want to be able to create an overlay of London with the data points. I also want to create a tutorial for people to be able to project location-based data points over other maps using Google Earth VR.

Risk: Every project has its risks. For one, I may not be able to overlay the map of London entirely as I might not have enough individual clusters. It also may lead to a more complex task then I thought if Google Earth VR won't allow me to properly overlay it. It should work, but the complexity may be more than I anticipate.

(3 hours)

Week 4: 16 Hours

2/13: Read more documentation about how to implement overlays on Google Earth VR (2 hours)

Read this super interesting article

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4648228/  (2 hours)

2/14 - Pre-Project self-evaluation (1 hour):

• The proposed project clearly identifies deliverable additions to our VR Software Wiki

  • 5 : I'm making a tutorial

• The proposed project will inform future research, ie, advancing human knowledge

  • 5 : The project will help people overlay data clusters over locations to better understand their data and why it clusters that way/

• The proposed project involves large data visualization along the lines of the "Data Types" wiki page and identifies the specific data and software that it will use

  • 5 : I have a large data set to use of orders and their location IDs

• The proposed project has a realistic schedule with explicit and measurable milestones at least each week and mostly every class

  • 4 : The milestones I've set are reasonable

• The proposed project includes an in-class activity

  • 5 : Yes, the game I proposed will be entertaining and insightful

• The proposed project has resources available with sufficient documentation

  • 5 : I have more than enough documentation to complete this project 

Read through current student research on the course wiki (3 hours)

2/16: Super interesting document to help me for the project (2 hours)

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4648228/ 

Made a lot of Progress on data clustering (3 hours) 

2/17: Completed first Milestone!  (3 hours)

Software I will use:

• Open Street Map works better than Google Earth VR (rolling with Open Street Map for Now) 

• https://www.kdnuggets.com/software/clustering.html for Data Clustering

• Created a list of steps to complete the project

• Almost finished with data clustering (part of Milestone 3)

Week 5: 10 Hours

2/19: Researched working with Open Street Map (2 hours)

https://www.youtube.com/watch?v=T9ap2lDPnJw

https://blog.mapbox.com/clustering-millions-of-points-on-a-map-with-supercluster-272046ec5c97

2/20: Went to Algorithmic Justice: Race, Bias, and Big Data talk and spoke to organizers afterwards

Super interesting talk!

(3 hours)

2/22: Worked on visualizing the data in 2D (4 hours)

2/24: Read through 2018 research findings in the wiki (2 hours)

Week 6: 10 Hours

2/26: 

Expiremented with Open Street Map to see how i could visualize my data. (6 hours)

2/28:

Expiremented with Google Earth VR to see how I could better visualize my data (valentine's day :( 4 hours)

Week 7: 10 Hours

3/3: Read a super interesting paper on VR and where its headed (2 hours)

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6232426/

3/5: Continued work on my project, attempted to solve map overlay on Google Earth VR (4 hours)

3/8: Created the 2d visualization of the data for my project, was super big so took a random sample of the largest clusters and put them on a smaller 

part of London. (4 hours)

Week 8: 10 Hours

3/10: Could not get the overlay working with Google Earth VR, worked on implementing it in Open Street Map (4 hours)

3/13: Really interesting paper on gaming in VR  (2 hours)

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0200724

3/14:

Project Status: 

I have decided on dropping Google Earth VR, as I have spent several hours attempting to add the overlays to Google Earth VR, but there is just no way to do it with the current system. It has been a huge annoyance, but I have to use Open Street Map. Doing this isn't hard, but it is nowhere near as useful as Google Earth VR. Below, I have images of what it looks like in Open Street Map (not happy with it).

I am going to change my project to become a software review of Google Earth VR and Open Street Map VR, and am going to make a tutorial on how to put Open Street Map in VR (3 hours of work).

I have completed all milestones that were needed prior to implementing it in Google Earth VR, but have changed my deliverables to a tutorial and software review.

3/16: Worked on Software Review (1 hour)

Week 9: 10 Hours

3/18: Worked on tutorial (3 hours)

3/21: Read through alot of the course wiki (3 hours)

3/25: Worked on tutorial and software review (4 hours)

Week 10: 10 Hours

First Project Results: 

• The Results of my first project were not very promising:

  • I originally wanted to make on overlay of London with my Quiqup data in VR

  • Google Earth didn't work, and Open Street Map was ugly and not very useful as a visualization

  • Instead, I made a tutorial on how to put Open Street Map in VR and a comparison between Google Earth VR and Open Street Map VR.

    • All in all, this second part was very successful. (4 hours working on it)

3/28: Read this article and researched software for surgery VR use cases

https://zspace.com/blog/Doctors-Use-VR (2 hours)

3/29: Read this article about VR usage for pain management! (1 hour)

https://health.usnews.com/health-care/patient-advice/articles/2019-01-14/how-virtual-reality-can-help-treat-chronic-pain

3/30: Worked on review for Google Earth vs Open Street Map VR (3 hours)

Week 11: 10 hours 

4/1: Presented project and continued to work on tutorial and review (4 hours)

4/3:

SECOND PROJECT DESCRIPTION:

Title - MRI Data Visualization

Description: Last Semester, I worked with Professor Badre in his lab conducting analyses on where clusters of memory are located in the brain. I wanted to model these clusters  in Paraview to be able to better understand if their was a correlation between the trials in the expirement, which showed images to different people and wanted to determine if those images were located in the same area of the brain as other people. 

Contributions: My main contribution is to create a visualization of the brain with the clusters inside using Paraview

Prior Knowledge: My work in the Badre Lab has helped me understand how voxels in the brain are grouped, and we have done a Paraview tutorial in class. 

Milestones:

• Week 1 

  • Cluster the data 

• Week 2 

  • Learn how to use Paraview better

• Week 3

  • Use Paraview to cluster the data

• Week 4

  • Use Paraview to cluster the data

• Week 5

  • Clean up the image

• Week 6

  • Write up documentation for others to be able to repeat it.

Deliverable: My project will have two concrete deliverables. First, I want to create the model in Paraview. The second is a tutorial for others to be able to do the same!

Risk: Every project has its risks. For one, the data may not cluster and just be a series of random points across the brain. That is my greatest concern, but regardless I will be able to produce a model

(4 hours)

4/5: Super interesting paper on VR work in agriculture (found it randomly online, but was captivated while reading it!) (2 hours)

Didn't think I'd enjoy a paper about agriculture so much!

https://link.springer.com/chapter/10.1007/978-3-642-12220-0_79 (2 hours)

Week 12: 10 hours

4/8: Worked on creating the pipeline to process the fMRI data (3 hours)

4/11: Completed pipeline, ran data processing, have the final CSV clusters (4 hours)

4/13: Read this paper on how to visualize fMRI data using Paraview (2 hours)

 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4648228/

Worked on implementing ideas in paper to visualize fMRI clusters (1 hour)

Week 13: 10 Hours

4/15:  Worked on visualizing the data clusters in Paraview (3 hours)

4/17: Finished visualizing the data clusters in Paraview, added a brain mask and condensed the clusters to fit inside the brain mask (4 hours (converting coordinates to brain mask took a verrrry long time.)

4/20: Worked on wiki contribution, where I made a tutorial describing how to visualize the clustets in Paraview. (3 hours)

Week 14: 10 Hours

4/22: Finished wiki contribution. It is now live under 2019 student research (3 hours)

4/23: Mid Project Checkin:

• Completed data clustering, most of wiki contribution is done.

• Need to improve the visualization, but it looks ok for this point in time.

4/25: Read a super interesting paper about object detection using neural networks (2.5 hours)

https://arxiv.org/abs/1809.03193

4/26: Read a really interesting paper about conversational AI using Neural Networks (2.5 hours)

https://arxiv.org/abs/1809.08267

4/28: Completed last touches on model, edited some of the visualization, cleaned up image and aninmation (2 hours)

Week 15: 10 hours

5/1: Worked on slideshow for in class presentation (2.5 hours)

5/2: Worked on poster for the demo day (2.5 hours)

5/4: Completed slideshow for in class presentation, completed all code and put it in wiki contribution. (2 hours)

5/7: Completed poster, did final preparation for demo day! (3 hours)

Journal reviewed by Andrej Simeski, updated according to his comments.