As a group, we are separating the goals of completing multiple models by groups influenced by an individual's learning style, knowledge, and times available.
Goal:
- To develop an understanding of how we can monitor earthquakes while efficiently opening an alarm monitoring system that identifies reduced costs and resources while maximizing effectiveness.
Models:
- Epidemic-Type Aftershock (ETAS) Model
- Magnitude Alarm Model
- Stark Model
Group Goals:
- Horizontal:
- To complete an analysis of all three models, ease communication between the Curators and Visualizers. With the Curators, we need to let them know we need a data set of earthquakes with all variables set up as a data frame associated with earthquakes. In our portion of the project, we will be manipulating the data to produce an analysis of the multiple models and set of information as an output for the visualizers to use. Within our group, three sub-groups have formed to accomplish certain goals.
- Group 1:
- To recheck our peers' work, maintain compatibility within each portion of our project, and coordinate between the Curators and Visualizers. Writing up the step-by-step instructions to complete the analysis properly and relaying necessary anticipated needs between groups will help to produce a reproducible set of code. Primarily working during the week and remotely as individuals.
- Group 2 & 3:
- A combination of individuals with extensive computer programming knowledge and analytical skills that are able to work on the weekends primarily as a group. Three models will be evaluated with the data frame given to Group 1 by the Curators.
Preparation for final project. Notes from office hours with Aaron Culich regarding the analyzer's portion of the final project.
Stages:
- Worry & Confusion
- Stochastic processes and depth of statistical knowledge
- How to communicate and coordinate among a group of 11 members
- Replicating Ph.D.'s work with various models
- Ambiguity & Questions
- What is analyzer role in the project
- We understand workflow and library files, but what work does that entail exactly to get there
- How many models are there (ETAS, poisson, simple Stark, etc.)
- Understanding inputs and outputs of each of the models
- Do 11 people all work on the analyzer role together, how do we break up the work
- Defined-ness
- Before and during Tuesday's class, we'll have defined roles for each person in the class, including analyzers
- Break out into groups of 2-5 to work on select parts of the assignment, either across horizontals or amongst the analyzers
- Be able to define and understand at a high level each of the models, how they differ
- Understand how to communicate with the other horizontal groups in terms of passing along information
- Prioritize
- After Tuesday's class, will have prioritized tasks for each individual
- S.M.A.R.T goals
- After Tuesday's class, have come up with S.M.A.R.T goals for groups and individuals and have an understanding each Tuesday of what we're presenting the following Tuesday in class
Diagramming the Workflow (see diagram below):
- Cloud of data from the real world
- Data from the Southern California Earthquake Center
- Other earthquake data available to the public
- Curation: Completed by the Administrators.
- Will cache the data, will curate the data
- Have a curated data in some format to pass along to analyzers
- The data will contain columns of variables that the analyzers may want to use to analyze
- Analyzers: Completed by the Producers.
- For the models, ask: how many models (i.e. 3?), what is the name of the model (i.e. ETAS, simple Stark, poisson), what is the math / statistics behind it, what are the inputs, what are the outputs
- For each model, understanding what the inputs and outputs are, bringing us closer to the defined-ness stage
- The members of the analyzer group can split into groups of 2-4 to each focus on a different model, or a different part of the workflow, such as a few focus on ETAS, a few focus on simple Stark, a few are communicators with the neighboring data curators and visualizers to understand inputs / outputs between groups; this will be figured out before and during class on Tuesday
- Analyzers should understand that output from data curators should be able to be used in any of the models that we choose (or after a bit more cleaning / analysis from our end), that our output from any model should be able to be used by the visualizers without much modification; any of the models that we use will likely ultimately come out with a single output to pass along to visualizers
- Visualizers: Completed by the Entrepreneurs.
- Models (i.e. a graph similar to curve in dissertation)
- Other visualizations
- Presenters: Performed by the Integrators.