How would you use Pandas to solve a data problem using groupby and merge, and what ML theory concepts would you connect it to?
Translate the data manipulation into PostgreSQL using the provided evaluation and project tables. Include completed evaluations with non-null actual and predicted values in the error metrics.
project_id, project_name, model_family, evaluation_count, mean_absolute_errorproject_id ascending| Column | Type | Description |
|---|---|---|
| project_idPK | INT | Unique project identifier |
| project_name | VARCHAR(100) | Project display name |
| model_family | VARCHAR(50) | Model family evaluated for the project |
| Column | Type | Description |
|---|---|---|
| evaluation_idPK | INT | Unique evaluation identifier |
| project_id | INT | Project associated with the evaluation |
| model_version | VARCHAR(30) | Version used to generate the prediction |
| status | VARCHAR(20) | Evaluation lifecycle status |
| actual_value | DECIMAL(10,2) | Observed target value |
| predicted_value | DECIMAL(10,2) | Model prediction |