NODE.md

Node Docs [F#]

How to Run

1. Prepare Observations File: Create an observations.json file with the following structure:

   [
     {
       "latitude": 34.052235,
       "longitude": 118.243683,
       "timestamp": 1622548800000
     },
     {
       "latitude": 40.712776,
       "longitude": 74.005974,
       "timestamp": 1622635200000
     },
     {
       "latitude": 51.507351,
       "longitude": 0.127758,
       "timestamp": 1622721600000
     }
   ]

2. Run Backend:

   cd node
   dotnet run

3. Run Frontend:

   cd frontend
   yarn install
   yarn dev

4. Load Observations in Browser:

   • Open your browser and navigate to http://localhost:3000.
   • Drop the observations.json file into the application.

5. Or Run Surprise Backtest Manually:

   curl -X POST http://localhost:5000/backtest/model-1 -H "Content-Type: application/json" -d @observations.json

   It will output the surprise for each observation, and the total model surprise over all observations:

   {
     "modelName": "model-1",
     "surprises": [
       {
         "latitudeSurprise": 61161.24970075105,
         "longitudeSurprise": 63742.415754987414,
         "timestamp": "2021-06-02T12:00:00",
         "totalSurprise": 124903.66545573846
       },
       {
         "latitudeSurprise": 63305.268561105564,
         "longitudeSurprise": 63937.053127532155,
         "timestamp": "2021-06-03T12:00:00",
         "totalSurprise": 252145.98714437618
       }
     ],
     "totalSurprise": 252145.98714437618
   }

Folders

• node: Contains the main program and related files for running the DeSciNet application.
• model-1: Includes the first version of the human movement model and its related files.
• model-2: Contains an alternative human movement model with different assumptions and structures.
• scm: Holds the structural causal model (SCM) framework and related utilities.
• google-timeline: Manages data export from Google location history.