Objective:

In this assignment, you will conduct a nowcasting analysis using the MIDAS (Mixed Data Sampling) approach to predict a policy indicator. You will leverage both low-frequency economic data and high-frequency data, including trends related to "deglobalization" from Google Trends. This assignment will require data handling, model building, and interpretative skills to assess forecasting performance and economic implications.

Instructions:

1. Data Loading:

o Low-Frequency Data: Use the provided monthly_indicators.xlsx file and load the data from the specified worksheet. Convert the date column to date format, and perform any necessary renaming of columns.

o High-Frequency Data: Download high-frequency weekly or daily data on the topic of "deglobalization" from Google Trends. You may include additional related keywords to enrich the dataset on "deglobalization." Make sure to preprocess this data for compatibility with the low-frequency data.

2. Data Processing:

o Correlation Analysis: Use feature selection techniques (such as correlation analysis) to identify and retain highly relevant variables.

3. MIDAS Model Construction and Nowcasting:

o Build a MIDAS model using the combined low-frequency economic indicators and high-frequency data from Google Trends.

o Experiment with different model configurations, recording model parameters and results for each configuration.

4. Results Visualization:

o Visualize the forecasting results, comparing model fit and predictive accuracy.

o Include a brief analysis explaining the economic interpretation of the results, the relevance of "deglobalization" trends, and any observed limitations of the models.

5. Submission Requirements:

o Submit your code files along with visualized output results.

o Provide a concise report that summarizes data preprocessing, model selection, forecasting results, and conclusions, with particular attention to the role of high-frequency "deglobalization" data in improving the model.