0.1.1 Cleaned up PairwiseCorrelation, matrix_print

0.1.2 Added JSON based examples

CHANGELOG.rst

@@ -3,6 +3,12 @@ ChangeLog
 
 PLEASE NOTE THAT THE API IS STILL UNSTABLE AS MORE USE CASES / FEATURES ARE ADDED REGULARLY
 
+v0.1.2 (26-03-2019)
+-------------------
+
+* Added example matrix_from_json_data
+* Cleaned up PairwiseCorrelation, matrix_print
+
 
 v0.1.0 (5-03-2019)
 -------------------

correlationMatrix/utils/preprocessing.py

@@ -44,13 +44,42 @@ def csv_files_to_frame(list, directory, filename):
  return df
 
 
-def construct_log_returns(in_filename, out_filename):
+def json_file_to_frame(input_filename, output_filename):
+ """
+ Given a file name with json data in the format
+ {
+ "Entity1" : [Values],
+ "Entity2" : [Values],
+ ...
+ "EntityN" : [Values]
+ }
+ Convert the data to a pandas dataframe for further processing
+
+ """
+
+ entity_data = pd.read_json(input_filename)
+ # select_data = entity_data.drop(columns=['High', 'Low', 'Open', 'Close', 'Volume'])
+ # index_data = select_data.set_index('Date')
+ # rename_data = index_data.rename(columns={"Adj Close": entry_name})
+ # df = pd.concat([df, rename_data], axis=1, sort=False)
+ #
+ entity_data.to_csv(output_filename, index=False)
+
+ return entity_data
+
+
+def construct_log_returns(in_filename, out_filename, drop_columns=None):
  """
  Load a dataframe with level data from file
+ Drop a list of columns that are not to be processed
  Store to file
 
  """
- level_data = pd.read_csv(in_filename).drop(columns=['Date'])
+ if drop_columns:
+ level_data = pd.read_csv(in_filename).drop(columns=drop_columns)
+ else:
+ level_data = pd.read_csv(in_filename)
+
  log_return_data = pd.DataFrame()
 
  for column in level_data:
@@ -67,9 +96,11 @@ def normalize_log_returns(in_filename, out_filename):
  Store to file
 
  """
+ mean_vals, std_vals = [], []
  log_return_data = pd.read_csv(in_filename)
  data = log_return_data.values
  cols = list(log_return_data)
+ print('Entity Names: ', cols)
  scaled_data = np.asarray(data)
  for ts in range(data.shape[1]):
  mean = data[:, ts].mean()
@@ -78,3 +109,9 @@ def normalize_log_returns(in_filename, out_filename):
 
  scaled_returns = pd.DataFrame(scaled_data, columns=cols)
  scaled_returns.to_csv(out_filename, index=False)
+
+ for ts in range(data.shape[1]):
+ mean_vals.append(scaled_data[:, ts].mean())
+ std_vals.append(scaled_data[:, ts].std())
+
+ return mean_vals, std_vals