Grouped descriptives

GroupedDescriptives.py

"""
Grouped descriptives
 
(c) 2013, Alan James Salmoni
"""
 
def Count(data):
    return len(data)
 
def Sum(data):
    val = 0
    for element in data:
        val = val + element
    return val
 
def Mean(data):
    return Sum(data) / float(Count(data))
 
class variable(list):
    def __init__(self):
        self.name = ''
 
def GetGroups(groupingvars):
    groups = []
    k = len(groupingvars)
    N = len(groupingvars[0])
    for idx in range(N):
        row = [var[idx] for var in groupingvars]
        if row not in groups:
            groups.append(row)
    return groups
 
def ExtractGroupsData(group, groupingvars, variable):
    N = len(variable)
    data = []
    for idx in range(N):
        row = [element[idx] for element in groupingvars]
        if group == row:
            data.append(variable[idx])
    return data
 
def GroupedDescriptives(groups, groupingvars, variables, stats):
    groupnames = ["Group 1","Group 2"]
    scorenames = ["Score 1","Score 2"]
    k = len(groups)
    table = '<h3>Describe</h3>\n<table>\n'
    for column in groupnames:
        table += '<th>%s</th>'%column
    for stat in stats:
        table += '<th>%s</th>'%(stat)
    table += '\n'
    table += '</table>\n'
    return table
 
def GroupDescriptives(groups, groupingvars, variables, stats):
    groupnames = ["Group 1","Group 2"]
    scorenames = ["Score 1","Score 2"]
    table = '<h3>Descriptives</h3>\n<table border="1">\n'
    table += '\t<tr><th>Variable</th>'
    for groupname in groupnames:
        table += '<th>%s</th>'%(groupname)
    for stat in stats:
        table += '<th>%s</th>'%(stat)
    table += '</tr>\n'
    k = len(groups)
    for idx, var in enumerate(variables):
        table += '\t<tr>'
        #for varname in scorenames:
        table += '<td rowspan="%d">%s</td>'%(k, scorenames[idx])
        for group in groups:
            for element in group:
                table += '<td>%s</td>'%element
            data = ExtractGroupsData(group, groupingvars, var)
            if "Count" in stats:
                val = Count(data)
                table += '<td>%s</td>'%str(val)
            if "Sum" in stats:
                val = Sum(data)
                table += '<td>%s</td>'%str(val)
            if "Mean" in stats:
                val = Mean(data)
                table += '<td>%s</td>'%str(val)
            table += '</tr>\n'
        table += '</tr>\n'
    table += '</table>\n'
    return table
 
 
if __name__ == "__main__":
    a = ["group 1","group 1","group 1","group 1","group 2","group 2","group 2","group 2"]
    b = ["Part 1","Part 1","Part 2","Part 2","Part 1","Part 1","Part 2","Part 2"]
    c = [1,2,3,2,1,3,2,4]
    d = [7,8,9,10,12,9,6,7]
    stats = ["Count","Sum","Mean"]
    groups = GetGroups([a,b])
    print GroupDescriptives(groups, [a,b], [c,d], stats)