Group 18: Shreyas Borse, I-Wei (Claire) Chen, Praharsha Mahurkar, Jeremy Wong

Cellular Key Performance Indicators (KPIs) Analysis and Comparison

Comparing the key performance indicators (KPIs) of a mobile network for different streaming services (Amazon Prime and Netflix) and mobility statuses (static and driving).

Dataset: GitHub: uccmisl / 5Gdataset

[TO-DO]

Regarding GitHub Repository - Well documented, reusable Python module targeted at solving your proposal task. Follow all good coding practices and conventions discussed in class

• Include all your code as .py files (therefore remember to write modular code to make life easier) 
• Include 1 Jupyter notebook which shows all the visualizations you generated for your presentation.
• Include a readme file that explains your file structure, how to run your code, and name all third-party modules you are using.
• Include the final presentation as a pdf file

Sample README Template

File Structure

.
├── README.md               <- README
│
├── 5Gdataset-master
│   │
│   ├── Amazon_Prime        <- Data collected using Amazon Prime 
│   │   │        
│   │   ├── Driving         <- Data collected while driving
│   │   │   │
│   │   │   ├── animated-AdventureTime      <- Data collected from watching animated show
│   │   │   │   ├── B_2019.11.28_07.27.57_cleaned.csv
│   │   │   │   ├── B_2019.11.28_07.27.57.csv
│   │   │   │   ├── ...
│   │   │   │   └── combined.csv            <- Cleaned and combined data from directory
│   │   │   │
│   │   │   └── Season3-TheExpanse      <- Data collected from watching filmed show
│   │   │       ├── B_2019.12.01_12.11.21_cleaned.csv
│   │   │       ├── B_2019.12.01_12.11.21.csv
│   │   │       ├── ...
│   │   │       └── combined.csv
│   │   │
│   │   └── Static          <- Data collected while stationary 
│   │       │                  (subdirectories follow same convention as Driving)
│   │       ├── animated-Ninjago
│   │       └── Season3-TheExpanse
│   │
│   ├── Download            <- Data collected when downloading content - not show-specific
│   │   ├── Driving
│   │   └── Static
│   │
|   └── Netflix             <- Data collected using Netflix 
|       │                      (subdirectories follow same convention as Amazon Prime) 
│       ├── Driving
│       │   ├── animated-RickandMorty
│       │   └── Season3-StrangerThings
│       │
│       └── Static
│           ├── animated-RickandMorty
│           └── Season3-StrangerThings
│
├── plots                   <- Plots created by data_plot.py
│   ├── Amaz_SvD_DL_bitrate.png
│   ├── ...
│   └── Stat_AvN_RSSI.png
│
├── data_cleanup.py         <- Script to clean-up and consolidate data within
│                              each subdirectory of 5Gdataset-master folder
│
├── data_plot.py            <- Script to plot data from certain combined.csv files
│
├── generated_visualizations.ipynb          <- Jupyter notebook that creates  
│                                              visualizations for presentation
└── [insert PDF file name here]             <- PDF of presentation

How to Run

Note: the output files (e.g. the cleaned data files and the plots) already exist in the repository, so the following steps are simply to describe the process as if they had not already been generated. If desired, you can remove the PNG files in the plots subdirectory, along with the _cleaned.csv and combined.csv files in each of the subdirectories of 5Gdataset-master.

First, navigate to the ECE143-Project directory.

Run data_cleanup.py to create cleaned and combined CSV files.

Following the previous step, there are two options for generating the plots:

1. Run data_plot.py to save all the plot figures into the plots subdirectory without showing the plots.
2. Run each cell in generated_visualizations.ipynb, starting with the first cell that contains from data_plot import plot_df, to show the plots in the output plots of each cell in the Jupyter Notebook (while also saving the plot figures into the aforementioned plots subdirectory).

If you wish to plot datasets or y-axis values not plotted in data_plot.py or generated_visualizations.ipynb, simply change or add the desired file paths at the top of these files and use as parameters to the plot_df function. Other parameters you can pass into plot_df are as follows:

• y_axis (str): the column name to use for the y-axis of the plot(s); default is DL_bitrate
• day (str): the day to use for the plot; default is 'Day1'; ignores this argument if subplots == True
• subplots (bool): plots only one plot if set to False, otherwise plots a 2x2 grid of subplots from Day 1 through Day 4 (assuming there are at least four days in the dataset); default is False
• show (bool): shows the figure if set to False, otherwise closes the figure; default is False
• columns (list(str)): the list of column names that will be accessed in the function call; default is ['Day','Timestamp','DL_bitrate','RSRQ','RSRP','RSSI']; requires 'Day', 'Timestamp', and y_axis at minimum

Third-Party Modules

• NumPy
• Pandas

Sources

D. Raca, D. Leahy, C.J. Sreenan and J.J. Quinlan. Beyond Throughput, The Next Generation: A 5G Dataset with Channel and Context Metrics. ACM Multimedia Systems Conference (MMSys), Istanbul, Turkey. June 8-11, 2020.