IDS Final Project: Pixlise+

Online URL: https://cmu-ids-2020.github.io/fp-airbnb-rats/

• Team members:
  • Constance Ye (constany@andrew.cmu.edu)
  • Lukas Hermann (lhermann@andrew.cmu.edu)
  • Nur Yildirim (nyildiri@andrew.cmu.edu)
  • Shravya Bhat (shravyab@andrew.cmu.edu)
• Track: Interactive Visualization/Application

Abstract

NASA JPL scientists working on the micro x-ray fluorescence (microXRF) spectroscopy data collected from Mars surface perform data analysis to look for signs of past microbial life on Mars. Their data analysis workflow mainly involves identifying mineral compounds through the element abundance in spatially distributed data points. Working with the NASA JPL team, we identified pain points and needs to further develop their existing data visualization and analysis tool. Specially, the team desired improvements for the process of creating and interpreting mineral composition groups. To address this problem, we developed an interactive tool that enables scientists to (1) semi-automatically cluster the data, and (2) compare the clusters and individual data points to make informed decisions about mineral compositions. Our tool supports a hybrid data analysis workflow where the user can manually refine the machine generated clusters.

Link to paper: In Markdown

Demonstration Video

Setting up the software

Frontend:

Required:

• Node

Installing packages:

npm i

Running:

npm run start

Building:

npm run build

Backend

Installing packages:

pip install -r requirements.txt

Running:

python3 backend2.py

If running your own server, be sure to change the backend URL in the Cluster.js file to the new url and port, for example: http://127.0.0.1:5000/

Clustering server rate limitations

We are running the clustering backend on an AWS instance. Because it's a free trail of an instance, it will expire around Janurary 9th, 2021. The app will continue to work afterwards, but the clustering feature will not work unless you set up your own server. We are also using a free proxy server service (CORS-anywhere) to forward our CORS requests because otherwise we will get https client to http errors. Due to the limits of free AWS and the free proxy server service, the rate and speed of responses may be limited if the app traffic is very high.

In this case there are two possible solutions:

1. Contact the team mebmers who can set up an alternative proxy server for you
2. Clone our repository and set it up locally using the instructions above.

Work distribution

Please see our task assignment spreadsheet for details on this.

Process

This application was developed over the course of about a month. We first had an initial meeting with one of our stakeholders, the project lead UX manager at NASA JPL where we gathered resources and learned about the possible problems that we could tackle. After more research into the datasets provided and some brainstorming about possible problem scopes, we presented initial mockups to the UX manager. Based on the feedback, we produced new mockups of 10 possible new interaction models or features and presented them to the co-investigator astrologists, who is a main beneficiary. Their input helped to validate our initial mockups, guide future design decisions and also helped us to prioritize which features to focus on. In the remaining weeks, we built the final application based on these findings.

Process mockups

Initial sketches First pass at showing comparison 10 possible interactions

Special thanks

Thank you to IDS professors Dominic Mortiz and Adam Perer and the TAs for their help!