Quick start

Tally is available on pip, to install it run

pip install datasmoothie-tally-client

import tally

Working with different data sources

Tally works with SPSS, CSV files, the Confirmit API and Unicom/Dimensions files (mdd/ddf). Here we using an SPSS file.

Note

You need a Tally API key to proceed, get in touch to get yours. Here, we store the tally key in an environment variable.

import os
dataset = tally.DataSet(api_key=os.environ.get('tally_api_key'))
dataset.use_spss('data/Example Data (A).sav')

Recode a variable

We will recode a variable called q14r06c03 which asks whether people agree that the waiting time in Store 3 is acceptible using Tally’s recode fucntion (documented here).

We want to reverse the answer codes so that 1 becomes 5, 2 becomes 4 etc.

First we take a look at what the meta currently looks like:

dataset.meta(variable='q14r06c03')

	codes	texts	missing
1	1	Strongly disagree	None
2	2	Disagree	None
3	3	Neither agree nor disagree	None
4	4	Agree	None
5	5	Strongly agree	None

The recode method can take very complicated mapping instructions, so when we recode we have to manage the meta data ourselves and make sure it matches the new encoding. So, we create a mapper for the labels as well

new_label_mapper = {(i+1):k for i,k in enumerate(reversed(list(dataset.meta(variable='q14r06c03')['texts'])))}
new_label_mapper

{1: 'Strongly agree',
 2: 'Agree',
 3: 'Neither agree nor disagree',
 4: 'Disagree',
 5: 'Strongly disagree'}

Then we create a mapper for the data and run recode and set_value_texts.

mapper = {
    5: {'q14r06c03':[1]},
    4: {'q14r06c03':[2]},
    3: {'q14r06c03':[3]},
    2: {'q14r06c03':[4]},
    1: {'q14r06c03':[5]}
}
dataset.recode(target='q14r06c03', mapper=mapper)
dataset.set_value_texts(name='q14r06c03', renamed_vals=new_label_mapper)
dataset.meta(variable='q14r06c03')

	codes	texts	missing
1	1	Strongly agree	None
2	2	Agree	None
3	3	Neither agree nor disagree	None
4	4	Disagree	None
5	5	Strongly disagree	None

dataset.crosstab(x='q14r06c03')

	Question	Total
	Values	Total
Question	Values
q14r06c03. Store 3 - The wait time when checking out was acceptable.	Base	4091.0
	Strongly agree	0.0
	Agree	1040.0
	Neither agree nor disagree	1007.0
	Disagree	1011.0
	Strongly disagree	1033.0

Removing answer codes

dataset.remove_values(
    name="q14r06c03",
    remove=[3]
)
dataset.crosstab(x='q14r06c03')

	Question	Total
	Values	Total
Question	Values
q14r06c03. Store 3 - The wait time when checking out was acceptable.	Base	3084.0
	Strongly agree	0.0
	Agree	1040.0
	Disagree	1011.0
	Strongly disagree	1033.0

Extend answer codes

dataset.extend_values(name='q14r06c03', ext_values=[[99, "No answer"]])
dataset.meta(variable='q14r06c03')

	codes	texts	missing
1	1	Strongly agree	None
2	2	Agree	None
3	4	Disagree	None
4	5	Strongly disagree	None
5	99	No answer	None