Skip to content

Hands-on Exercise

The exercise applies the full workflow to Philippines. Assume we are doing the October 2026 first round forecast.

Begin: Follow the steps below

Step 1. Update data

Run datastore_pumps.update_all to refresh the datastore (see Forcasting Flow: Update Data).

python -m datastore_pumps.update_all

Step 2. Forecast external economies.

Use this sample external economies forecast to tune the annual GDP growth.

Economy GDP growth 2026 GDP growth 2027
US 2.1 2.0
PRC 4.6 4.5
Euro area 0.8 1.2
Japan 0.7 0.8

United States (us) forecast

  1. Open and update the us/attributes.json

    {
        "forecast_id": "EX_OCT2026_F1",
        "forecast_end_period": "2027-Q4",
        "history_end_period": "2026-Q2",
        "history_start_period": "2010-Q1",
        "country_code": "US",
        "country_name": "United States"
    }
    
  2. Quick run all required steps

    python -m us.run_all
    
  3. Update the GDP annual tunes

    1. In the windows explorer, open annual_tunes.csv (path = c:/CPM_Tool/us/tune_files/).

    2. In the roc4_ann_gdpw column, enter 1.021 (2026) and 1.020 (2027).

    3. Save and close the tune csv file.

  4. Run python -m us.simulate_forecast, to compute the forecast.

  5. Run python -m us.report_forecast, to generate the charts. Review the chart, once ok, proceed to the next step.

  6. Run python -m us.submit_forecast, to save the US forecast to the datastore/data.csv.


PRC (cn) forecast

  1. Open and update the cn/attributes.json

    {
        "us_forecast_id": "EX_OCT2026_F1",
        "forecast_id": "EX_OCT2026_F1",
        "forecast_end_period": "2027-Q4",
        "history_end_period": "2026-Q2",
        "history_start_period": "2010-Q1",
        "country_code": "CN",
        "country_name": "China"
    }
    
  2. Quick run all required steps

    python -m cn.run_all
    
  3. Update the GDP annual tunes

    1. In the windows explorer, open annual_tunes.csv (path = c:/CPM_Tool/cn/tune_files/).

    2. In the roc4_ann_gdpw column, enter 1.046 (2026) and 1.045 (2027).

    3. Save and close the tune csv file.

  4. Run python -m cn.simulate_forecast, to compute the forecast.

  5. Run python -m cn.report_forecast, to generate the charts. Review the chart, once ok, proceed to the next step.

  6. Run python -m cn.submit_forecast, to save the PRC forecast to the datastore/data.csv.

Repeat the process for euro area (ea) and Japan (jp).


Step 3. Forecast Philippines

  1. Open and update the ph/attributes.json

    {
        "external_forecast_id": "EX_OCT2026_F1",
        "forecast_id": "PH_OCT2026_F1",
        "forecast_end_period": "2027-Q4",
        "history_start_period": "2010-Q1",
        "country_code": "PH",
        "country_name": "The Philippines"
    }
    
  2. This time, we will go through the nine steps. Run the commands one at a time.

    Hnads-on

    # First, configure forecast attributes in attributes.json
    python -m ph.create_model
    python -m ph.request_data
    python -m ph.prepare_history
    
    python -m ph.report_steady_calibration
    #  review calibration, then continue
    
    python -m ph.filter_history
    
    python -m ph.report_filter
    #  review filter results
    
    python -m ph.simulate_forecast
    #  apply judgmental adjustments
    
    python -m ph.report_forecast
    #  review forecast results, then submit
    
    python -m ph.submit_forecast
    

Other focus economies

The same sequence applies to any country module — replace ph with the country code, for example in for the India.