RooMBTI

2023 Spring, CS481 Data Visualization

Data Visualization to choose the best random roommate who has the most similar routines with me.

Team Information

Student ID	Name	GitHub ID
20160381	Seungjun Oh (오승준)	sjuuun
20170152	Jeongmin Kim (김정민)	sillysillyman
20190419	Junseok Youk (육준석)	lsz4221

Project Information

• Process Book
• Prototype URL

Installation

Install using requirements.txt

$ pip install -r requirements.txt

Following libraries are needed.

pandas==2.0.1
dash==2.9.3
dash-core-components==2.0.0
dash-bootstrap-components==1.4.1
plotly==5.14.1
notebook==6.5.4
gunicorn==20.1.0
scikit-learn==1.2.2

Directory Information

assets

• Assets such as css files.

csv

• Processed data for visualization.
• routines
  • Aggregated routines per user.
  • There are one csv file for daily routines and one csv file for weekly routines per one user.
  • {USER ID}-daily.csv
  • {USER_ID}-weekly.csv
• routines_raw
  • Aggregated routines.
  • {USER_ID}-location.csv
    • Aggregated routines from LocationEntity data.
    • It contains routines for indoor time, class time, study time, exercise time.
  • {USER_ID}-sleep.csv
    • Aggregated routines from DeviceEventEntity data.
    • It contains routines for sleep time.
    • If there are no event during more than 4 hours, we assume it to sleep time.
• bfi.csv
  • Data of personal traits from BFI-15 data.
• daily_routines.csv
  • Merged data of daily routines in routines directory.
• weekly_routines.csv
  • Merged data of weekly routines in routines directory.
• similarity.csv
  • Data for similarity of routines.

data_process

Jupyter Notebook or Python files for data processing.

figures

Methods for visualizing data with Plotly.

pages

Python files for page layout with Dash. Here, the Dash application calls figure functions in figures directory.

app.py

Main file of the web application. Run the web application by the following command.

$ python app.py

Configuration files

• .gitignore: List of files which are ignored in git repository.
• .mapbox_token: API token for visualization of Mapbox scatter plot.
• requirements.txt: List of required libraries.
• Procfile: Configuration for Heroku. (Command to be executed in Heroku)
• runtime.txt: Configuration for Heroku. (Python runtime version)

Pages

Following files are in pages directory. This is brief information about those files.

• overview.py
  • Overview of my routines.
  • User can analyze their own routines in this page.
• users.py
  • List of roommates.
  • User can compare roommates by similarity of sleep routines.
• comparison.py
  • Comparison between my routines and roommate's routines.
  • If user select one of roommates, user can compare routines with he/she.

Figures

Following files are in figures directory. This is brief information about those files.

• bfi.py: Pentagon widget for BFI-15 data.
• daily_routine.py: Bar chart of daily routine.
• weekly_routine.py: Bar chart of weekly routine.
• indoor.py: Line chart to visualize when users are in indoor.
• location_mapbox.py: Scatter chart to visualize location information.

Data Process

Following files are in data_process directory. This is brief information about those files.

• location.ipynb
  • Analyze routines by LocationEntity data.
  • We used DBSCAN to cluster geographical data and derive routines from location data.
  • Derive routines of indoor time, class time, study time, exercise time.
• sleep_by_device.ipynb
  • Analyze sleep routine by DeviceEventEntity data.
  • We assume that if device is not used more than 4 hours, it will be sleep routine.
• aggregate_routine.ipynb
  • Aggregate routines from the result of location and sleep data.
  • Aggregate into daily routines and weekly routines.
• concat_routines.py
  • Concat daily and weekly routines of each users.
• bfi.py
  • Select required data from BFI-15 data.
• get_similarity.ipynb
  • Calculate similarity of sleep routines.