# Simple data analysis for forecasting

This Python script extracts data from txt files containing raw numerical data, notably the increase average, the relative evolution and the standard deviation of the values. All of these values require a period on which they are calculated. This period is set via the command line.

#! /usr/bin/env python3

import statistics
import sys

if len(sys.argv) != 2:
print("""SYNOPSIS
./%s period

DESCRIPTION
period the number of days defining a period""" % (sys.argv[0]))
exit(1)
increases = []
period = int(sys.argv[1])
values = []

def check_length(array):
if len(array) > period:
array.pop(0)

def mean(data):
"""Return the arithmetic mean of data."""
return round(sum(data) / len(data), 2)

def calculate_g():
increase = 0
try:
increase = (values[-1] * 10000 - values[-2] * 10000) / 10000
except IndexError:
pass
if increase > 0:
increases.append(increase)
else:
increases.append(0)
if len(values) < period + 1:
print("g=nan ", end="")
else:
increases_average = mean(increases[1:])
print("g=%.2f " %  (increases_average), end="")

def calculate_r():
if len(values) < period + 1:
print("r=nan% ", end="")
return "nan"
else:
evolution = round((values[period] - values[0]) * 100 / values[0])
print("r=%.0f%% " %  (evolution), end="")
return (1, -1)[evolution < 0]

def _ss(data):
"""Return sum of square deviations of sequence data."""
average = mean(data)
ss = sum((x-average)**2 for x in data)
return ss

def stddev():
"""Calculates the population standard deviation by default."""
n = len(values)
if n < period:
print("s=nan", end="")
else:
ss = _ss(values[-(period):])
print("s=%.2f" % (round((ss / period) ** 0.5, 2)), end="")

def main():
line = input()
prev_sign = 0
say = ""
switch = 0
while line != "STOP":
values.append(float(line))
calculate_g()
sign = calculate_r()
stddev()
if sign != "nan":
if sign == prev_sign * -1:
say = " a switch occurs"
switch += 1
else:
say = ""
prev_sign = sign
print("%s" % (say))
check_length(values)
check_length(increases)
line = input()
print("STOP\nGlobal tendency switched %i times" % (switch))

if __name__ == "__main__":
main()


and here is sample data:

27.7
31.0
32.7
34.7
35.9
37.4
38.2
39.5
40.3
42.2
41.3
40.4
39.8
38.7
36.5
35.7
33.4
29.8
27.5
25.2
24.7
23.1
22.8
22.7
23.6
24.3
24.5
26.7
27.0
27.4
29.8
29.4
31.5
29.6
29.8
28.9
28.7
27.2
25.7
26.0
25.2
21.6
20.3
21.1
20.4
19.8
19.1
19.6
21.2
21.0
21.4
24.0
25.5
25.5
26.4
29.4
32.1
31.4
32.3
35.2
38.3
36.6
38.4
39.9
40.5
39.4
39.0
40.5
42.1
38.7
37.5
38.1
36.5
35.4
STOP


I'm pretty sure my code isn't that clean since I'm pretty new to Python and I would like to know good practices.

extracts data from txt files

but, the code does not read a file. It takes input from the standard input (stdin).

When I run the code, I see output such as:

g=1.00 r=25% s=0.00


You should explain in the comments what g, r and s mean.

It is good that the code has usage information, but you should add more information about the required period argument and how it is used in the code.

The code uses input to accept user input from stdin while running, but the user does not know what input is valid.

## Input checking

If the user inputs a string instead of a number, the code dies. Consider exiting more gracefully. For example, the special string STOP ends the run gracefully. Perhaps allow it to end when the user enters any non-numeric input.

## Names

You gave meaningful names to some of the variables and functions. However, some others could be improved. Consider:

def check_length(array):


The name array is too generic. check is good, but length is again too generic.

calculate_g: replace g with something more meaningful. The same for r in calculate_r.

## Lint check

I used pylint on the code, and I got this result (among others):

W0611: Unused import statistics (unused-import)


It is a good idea to remove unused code.

## Parsing command line

Consider using argparse instead of sys.argv for parsing the command line.

Here is the code with some of the above suggestions:

'''
Calculate statistics.

Input numbers from the command line.
'''

import sys

if len(sys.argv) != 2:
print("""SYNOPSIS
./%s period

DESCRIPTION
period the number of days defining a period""" % (sys.argv[0]))
exit(1)

increases = []
period = int(sys.argv[1])
values = []

def check_length(array):
if len(array) > period:
array.pop(0)

def mean(data):
"""Return the arithmetic mean of data."""
return round(sum(data) / len(data), 2)

def calculate_g():
increase = 0
try:
increase = (values[-1] * 10000 - values[-2] * 10000) / 10000
except IndexError:
pass
if increase > 0:
increases.append(increase)
else:
increases.append(0)
if len(values) < period + 1:
print("g=nan ", end="")
else:
increases_average = mean(increases[1:])
print("g=%.2f " %  (increases_average), end="")

def calculate_r():
if len(values) < period + 1:
print("r=nan% ", end="")
return "nan"
else:
evolution = round((values[period] - values[0]) * 100 / values[0])
print("r=%.0f%% " %  (evolution), end="")
return (1, -1)[evolution < 0]

def _ss(data):
"""Return sum of square deviations of sequence data."""
average = mean(data)
ss = sum((x-average)**2 for x in data)
return ss

def stddev():
"""Calculates the population standard deviation by default."""
n = len(values)
if n < period:
print("s=nan", end="")
else:
ss = _ss(values[-(period):])
print("s=%.2f" % (round((ss / period) ** 0.5, 2)), end="")

def main():
line = input('Enter a number or STOP to finish: ')
prev_sign = 0
say = ""
switch = 0
while line != "STOP":
values.append(float(line))
calculate_g()
sign = calculate_r()
stddev()
if sign != "nan":
if sign == prev_sign * -1:
say = " a switch occurs"
switch += 1
else:
say = ""
prev_sign = sign
print("%s" % (say))
check_length(values)
check_length(increases)
line = input('Enter a number or STOP to finish: ')
print("STOP\nGlobal tendency switched %i times" % (switch))

if __name__ == "__main__":
main()