1 files changed, 114 insertions, 0 deletions
diff --git a/datasets/construction/convert.py b/datasets/construction/convert.py
new file mode 100644
index 0000000..0abde23
--- /dev/null
+++ b/datasets/construction/convert.py
@@ -0,0 +1,114 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+"""
+Dataset conversion script
+
+Author: G.J.J. van den Burg
+
+"""
+
+import argparse
+import json
+import xlrd
+
+MONTHS = {
+    "Jan": 1,
+    "Feb": 2,
+    "Mar": 3,
+    "Apr": 4,
+    "May": 5,
+    "Jun": 6,
+    "Jul": 7,
+    "Aug": 8,
+    "Sep": 9,
+    "Oct": 10,
+    "Nov": 11,
+    "Dec": 12,
+}
+
+
+def format_date(datestr):
+    """ expects: mmm-yyx with x an extraneous character or empty """
+    mmm, yyx = datestr.split("-")
+    midx = MONTHS[mmm]
+    if len(yyx) == 3:
+        yy = yyx[:2]
+    elif len(yyx) == 2:
+        yy = yyx
+    else:
+        raise ValueError
+
+    # this will break in 71 years
+    if yy.startswith("9"):
+        yyyy = 1900 + int(yy)
+    else:
+        yyyy = 2000 + int(yy)
+    return f"{yyyy}-{midx:02}"
+
+
+def parse_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("input_file", help="File to convert")
+    parser.add_argument("output_file", help="File to write to")
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+
+    wb = xlrd.open_workbook(args.input_file)
+    ws = wb.sheet_by_index(0)
+    header = ws.row(3)
+    assert header[0].value == "Date"
+
+    by_month = {}
+    ridx = 4
+    while True:
+        # stop if date cell is empty
+        if ws.row(ridx)[0].ctype == xlrd.XL_CELL_EMPTY:
+            break
+
+        date_value = ws.row(ridx)[0].value
+        construct_value = ws.row(ridx)[1].value
+
+        date = format_date(date_value)
+        construct = int(construct_value)
+
+        by_month[date] = construct
+        ridx += 1
+
+    name = "construction"
+    longname = "US Construction Spending"
+    time = sorted(by_month.keys())
+    time_fmt = "%Y-%m"
+    values = [by_month[t] for t in time]
+
+    series = [
+        {
+            "label": "Total Private Construction Spending",
+            "type": "int",
+            "raw": values,
+        }
+    ]
+
+    data = {
+        "name": name,
+        "longname": longname,
+        "n_obs": len(time),
+        "n_dim": len(series),
+        "time": {
+            "type": "string",
+            "format": time_fmt,
+            "index": list(range(len(time))),
+            "raw": time,
+        },
+        "series": series,
+    }
+
+    with open(args.output_file, "w") as fp:
+        json.dump(data, fp, indent="\t")
+
+
+if __name__ == "__main__":
+    main()