integrate data quality measurements

[mjstealey]

@arunacs - I don't fully know what this will look like yet, but wanted a placeholder issue.

[arunacs]

Identified four specific bias/error metrics that we will compute for O3, NO2, SO2, CO, PM2.5 (and chemical components) for various observation networks in the country, and report domain-average hourly metrics for 2010 and 2011.

[mjstealey]

Updated reference

[mjstealey]

Proposal:

• Adding stanza to each value of cmaq_output for variables that have a data quality metric associated to them.
  • If the variable does not have an associated metric, then the empty set is returned "data_quality": {}

Example o3 output:

{
  "values": [
    {
      "variable": "o3",
      "lat_lon": "35,-80",
      "cmaq_output": [
        {
          "date_time": "2010-01-01 00:00:00",
          "value": 49.8195953369141,
          "data_quality": {
            "obs_average": 15.7729,
            "model_average": 27.5666,
            "bias_average": 11.7936,
            "rmse_average": 18.6757,
            "corr_average": 0.313
          }
        }
      ]
    }
  ]
}

[mjstealey]

Will add a data_quality boolean flag to the exposure_list table:

Table exposure_list:

- id, 
- variable, 
- description, 
- units, 
- common_name, 
- utc_min_date_time, 
- utc_max_date_time, 
- resolution, 
- aggregation, 
- data_quality

And then expand the data_quality table to include the columns for all variable metrics.

Table data_quality:

- id, 
- utc_date_time, 
- o3_model_average, 
- o3_obs_average, 
- o3_corr_average, 
- o3_rmse_average, 
- o3_bias_average, 
... (and so on)

The data_quality stanza from above will then be injected into the return value based on the boolean from exposure_list on query.

[mjstealey]

Code has been updated and merged into master.

Quality metrics are now made available in the /variables endpoint if they exist. Example for o3:

{
      "aggregation": [
        "max", 
        "avg"
      ], 
      "common_name": null, 
      "description": "1000.0*O3[1]", 
      "end_date": "2011-02-01 01:00:00", 
      "has_quality_metric": true, 
      "quality_metric": [
        {
          "common_name": "root mean squared error", 
          "variable": "rmse"
        }, 
        {
          "common_name": "mean observed value ", 
          "variable": "obs_mean"
        }, 
        {
          "common_name": "mean modeled value", 
          "variable": "mod_mean"
        }, 
        {
          "common_name": "mean bias", 
          "variable": "mb"
        }, 
        {
          "common_name": "pearson correlation coefficient", 
          "variable": "cor"
        }
      ], 
      "resolution": [
        "hour", 
        "day", 
        "7day", 
        "14day"
      ], 
      "start_date": "2010-12-01 00:00:00", 
      "units": "ppbV", 
      "variable": "o3"
    }

If a quality metric exists for a queried variable, it will be returned as part of of the /values output.

{
  "values": [
    {
      "cmaq_output": [
        {
          "date_time": "2010-12-30 05:00:00", 
          "quality_metric": {
            "cor": 0.426, 
            "mb": 11.4464, 
            "mod_mean": 35.0096, 
            "obs_mean": 23.5632, 
            "rmse": 12.9588
          }, 
          "value": 30.7911243438721
        }, 
        {
          "date_time": "2010-12-31 05:00:00", 
          "quality_metric": {
            "cor": 0.465, 
            "mb": 10.4739, 
            "mod_mean": 30.5484, 
            "obs_mean": 20.0744, 
            "rmse": 12.359
          }, 
          "value": 36.8378486633301
        }, 
        {
          "date_time": "2011-01-01 05:00:00", 
          "quality_metric": {
            "cor": 0.577, 
            "mb": -0.70803, 
            "mod_mean": 22.361, 
            "obs_mean": 23.069, 
            "rmse": 7.62782
          }, 
          "value": 21.735221862793
        }, 
        {
          "date_time": "2011-01-02 05:00:00", 
          "quality_metric": {
            "cor": 0.633, 
            "mb": -1.15105, 
            "mod_mean": 21.2295, 
            "obs_mean": 22.3806, 
            "rmse": 7.97999
          }, 
          "value": 18.5885848999023
        }, 
        {
          "date_time": "2011-01-03 05:00:00", 
          "quality_metric": {
            "cor": 0.438, 
            "mb": -0.191461, 
            "mod_mean": 19.4314, 
            "obs_mean": 19.6229, 
            "rmse": 8.51403
          }, 
          "value": 23.6635990142822
        }
      ], 
      "lat_lon": "35,-80", 
      "variable": "o3"
    }
  ]
}