Modeling People Movement

OVERVIEW: This project takes a dynamical systems approach to modeling the flow of people on a public campus. Everything is explained in the Jupyter Notebooks and will be updated over the next few months.

Workflow

1. Processing Procedure

   • Dataset Introduction: Campus WiFi
   • The DataProcessor Class
   • Class Methods
   • Optional Parameters
   • Speed-up: Parallel Data Processing
   • Performance Testing and Results

2. Data Visualization

   • Time-Series Visualization
   • The DataVisualization Module
   • Classes in this Module
     • The DataVisualizer Class
     • The InteractiveVisualizer Class

3. Matrix Decompositions

   • SVD
   • Truncated-SVD
   • NMF
   • TSNMF

4. Sparse Indentification of Nonlinear Dynamics

Directory Information

/ModelingPeopleFlow
    .env
    .vscode

    /data
        /images
            /3d
            /matrix-fact
            /speed-up
            /time-series
            /wordcloud

        /input
            /configs
            /coords
                campusCoords.json
                engineeringCoords.json

            /events
                event_dict.py  

            /png_files
                main_campus.png
                east_campus.png
                engineering_quad.png

            /WiFiData-old
                /Eduroam
                /UCBGuest
                /UCBWireless

            /WiFiData-new

        /output
            /building-plots
                /all-buildings
                    sparsity_3networks-AllData.png

                /chunks
                    /normalized
                        "%b-%d-%Y"_to_"%b-%d-%Y"
                            /svd
                            /trunc-svd
                            /nmf
                            /tsnmf

                    /special
                    /time-series

                /single
                    /normalized
                    /time-series
                    /interval

            /campus-plots
        /references
        /tables
        README.md
    
    /notebooks
        DataAnimator.ipynb
        DataDecompositions.ipynb
        DataVisualization.ipynb
        MAIN.ipynb
        TEST.ipynb
                    
    /src
        /python
            /data_plotter
                sparsity_plotter.py
                bar_charts.py

            /decomps
                nmf.py
                svd.py

            /processor
                config_manager.py
                data_processor.py
                sparsity.py
            
            /sindy

            /time_series
                creator.py
                event_plotting.py
                plot_builder.py

            /gui
                launcher.py

    .gitignore
    config.json
    config2.json
    README.md
    requirements.txt

Acknowledgement

Built as a side project during summer 2023 by Tyler Reiser. This code has never been run on CU Research Computing resources - the goal was to build a system that could process WiFi information reports for the entire campus, eventually running it continuously on the Alpine supercomputing cluster.

I, Tyler A. Reiser, acknowledge the support from multiple parties. From the University of Colorado Boulder, Office of Information Technology, I thank Glenn Rodrigues, Chief WiFi Network Arcitect, for providing access to this dataset. Not only was this a unique opportunity to work with a dataset relevant to the university, it provided access to data that almost no other researchers had access to. This would not have been possible without his approval. From the Department of Veteran Affairs, I thank Vicki George for unwavering support and, of course, overseeing the education benefits program that funded my graduate studies. These benefits gave me the freedom to explore the ideas that led to the development of this software and the discoveries that followed suit. Thank you.

---

CREATED: SEPTEMBER 2023
MODIFIED: NOVEMBER 2024

COPYRIGHT (c) 2024 Tyler A. Reiser