7:00 – 6:00 ASRC MKT

• Renting a spec Miata at Summit Point 
• This is really good: The Human Strategy A Conversation With Alex “Sandy” Pentland [10.30.17]
  • Human behavior is determined as much by the patterns of our culture as by rational, individual thinking. These patterns can be described mathematically, and used to make accurate predictions. We’ve taken this new science of “social physics” and expanded upon it, making it accessible and actionable by developing a predictive platform that uses big data to build a predictive, computational theory of human behavior.
• Rerunning the DTW with the selected agent weight being the specified weight rather than scaled by the distance from the angle so that it matches better the RANDOM_AGENT and the RANDOM_AGENTS settings.
• Ok, here’s the results. The relationships between the populations appears more consistent, but that could be normal variability. Time for some true statistics to see if these are actually distinct populations. I can also increase power by doing more runs. Possibly also increasing the population size, though there might be confounding effects. 
• Pandas can read in a specific Excel sheet and numpy can run bootstrap on DataFrames, so I can automate the analysis. Going to talk to Aaron first, since he might be the one to go down this road.
• I think the next step is to start on the UI for the polarization game. Angular?
  • • Installing NodeJS
    • npm install -g @angular/cli -> added 968 packages in 56.599s. That is a lot of packages. The IntelliJ plugin seems to be working, the @angular/cli package is visible: 
    • Creating a new project is reasonable 
    • Once the project is running, the way to compile and run seems to be to run ng serve –open in the IntelliJ terminal (Note: When running as non-admin, do this in a terminal with admin privileges). It then does a whole bunch of things when a code change is made:

      ** NG Live Development Server is listening on localhost:4200, open your browser on http://localhost:4200/ **
       10% building modules 8/10 modules 2 active ...\PolarizationGameOneUI\src\styles.csswebpack: wait until bundle finished: /                                                              Date: 2017-11-07T15:50:25.164Z
      Hash: b3174f5198d14bdc05ac
      Time: 4708ms
      chunk {inline} inline.bundle.js (inline) 5.79 kB [entry] [rendered]
      chunk {main} main.bundle.js (main) 20.8 kB [initial] [rendered]
      chunk {polyfills} polyfills.bundle.js (polyfills) 553 kB [initial] [rendered]
      chunk {styles} styles.bundle.js (styles) 33.8 kB [initial] [rendered]
      chunk {vendor} vendor.bundle.js (vendor) 7.02 MB [initial] [rendered]
      
      webpack: Compiled successfully.
      webpack: Compiling...
      Date: 2017-11-07T15:51:07.132Z
      Hash: 7b89b5a301e4a411e92d
      Time: 703ms
      

    • Everything is then sent to localhost:4200/, so all the browser debuggers are available
    • 
    • And you can change the picture in the app.component.html file. re-renders on the fly. Pretty nifty. Yep verified:The ng serve command builds the app, starts the development server, watches the source files, and rebuilds the app as you make changes to those files.The --open flag opens a browser to http://localhost:4200/.
    • Pleasantly, if the install fails, ng serve –open will complete the install nd then start the server.
    • Added the ‘heroes’ component:  
    • Then I got this error message:

      ERROR in src/app/heroes/heroes.component.ts(7,18): error TS2304: Cannot find name 'ViewEncapsulation'.

    • Turns out that I had to add ViewEncapsulation to the imports in heroes.components:

      import {Component, OnInit, ViewEncapsulation} from '@angular/core';
      
      @Component({
        selector: 'app-heroes',
        templateUrl: './heroes.component.html',
        styleUrls: ['./heroes.component.css'],
        encapsulation: ViewEncapsulation.None
      })
      export class HeroesComponent implements OnInit {
        constructor() { }
        ngOnInit() {
        }
      }

      Once added in, the rebuild happened and everything functioned normally. Correct error message in the IDE and everything!

• Talked to Aaron about next steps with the herding data. We need to do something with NNs, and this could be a good fit
• And now I have a nice little certificate of candidacy!

7:00 – 4:00 ASRC MKT

• Going to try a batch job that runs the sim on a single population with a .2 radius and see if I can see a difference between the behaviors using DTW.
• I had created a few bugs with changing the names of the flocks to Red and Green. Also, I had never run in batch mode with StorageAndRetreival. And calculations for an average center don’t work when there are no members of your flock. So fixing bugs.
• First set of outputs from the batch jobs. Here’s the headings: 
• And here’s the DTW for the same settings (smaller stage though for proportionally greater differences): 
• The first really obvious thing it that NoHerding is distinct from the other settings, which are more like echo chambers. Groupings tighten up as the radius increases, and the average heading approach may be statistically better than the random agents, but not by much. Lastly, RANDOM_AGENTS and RANDOM_AGENT lie on a continuum. As the switch between each agent takes longer, the more AGENTS will start to look like AGENT.