Time series analysis of CMPC data.
To install necessary requirments. ''' pip install -r requirements.txt ''' Rerun this command periodically to ensure versions are compatible with the latest version of this package.
Once requirements are installed, you may test the installation of streamlit using 'streamlit hello' .
This package contains two interfaces: a command line app and a web app. The command line app is in maintenance only mode. To run, navigate to the root directory for the package in the terminal and run 'python command_line_app.py --help' for instructions.
To run the web interface, use ''' streamlit run web_app.py '''
In both cases, input files must be of the appropriate format chosen, as described in inputs.
The package accepts the following input formats for both endogeneous and exogeneous data.
Endogeneous data:
excel_table: - xlsx file of only one sheet - Cell 1A must contain the term "Year" - Column A contains the year of the observation(s) in YYYY format. Must contain year label for every nonempty row. - Row 1 contains the month of the observations. May be in any format recognized by dateparser library. Not case sensitive. Must contain month label for every nonempty column. - There must be exactly one observation per month between the first and last months given, inclusive.
cmpc_financial_data: - csv file - "Description" appears in row 1. It's column contains the names of all variables, including endogeneous_var. It's row contains the quarter for each column of observations in the format QX YY. All other columns not in that format will be dropped. - There is exactly 1 column per quarter. There is exactly one column per quarter for each quarter between the min and max quarters specified. - endogeneous_var must appear in the column 'Description' exactly once. It's row contains the value for the endogeneous var. - endogeneous_var cannot be the empty string - '', '-', ' ' are treated as null values, in addition to NaN's. All nulls in quarterly columns are filled with 0.0's. - Blank rows are skipped - Negative numbers are represented with -, as in -123.45. () for negative numbers may NOT be used. Commas cannot be used to break apart large numbers (ie 123,456 is not allowed).
Exogeneous data: macro: - xlsx file - Data must be located in the only table on a sheet named 'BBG P' - Row 4 must contain unique column names, each representing a unique variable. - Rows 1, 2, 3, 5, and 6 are extraneous header information and will be skipped upon import. Row 7 contains the first row of actual data. Any rows with a missing index in column A are skipped. - Cell 4A must be empty or null. - Column A contains the date index for the data in '%m/%d/%y' format. - Data begins in column B. - Null values or empty strings are ok in data-containing cells, but will be backfilled with the least recent non-null cell. If a column is missing the most recent observation and cannot be backfilled, it will be forward filled. Completely empty columns are dropped.