Generate Automated Survey Reports with Python in Minutes

Solution Overview

Tool package: https://github.com/gasongjian/reportgen

Project address: https://github.com/gasongjian/ (star or fork)

Software dependencies: Python 3 (compatible with Python 2, but Chinese part needs a manual patch)

Data requirements: survey data from Wenjuanxing, Wenjuanwang, etc.

Main feature 1: Automatically generate overview reports (PPT) with frequency statistics and charts for each question.

Main feature 2: Automatic cross‑analysis reports (chi‑square test, TGI, CHI metrics, simple conclusions).

Preparation

Required environment:

Install Anaconda (recommended Python 3) for scientific packages.

Install third‑party package python‑pptx: pip install python-pptx Install the report package by placing report/report.py in your working directory or site‑packages.

Note

Python 2.7 version of python‑pptx has a Chinese‑font bug; edit .\\pptx\\chart\\xmlwriter.py lines 1338 and 1373, replace escape(str(name)) with escape(unicode(name)).

Quick Start

For users who do not know or do not want to learn Python 3, see this.

For Windows users, download the reportgen folder from the project and follow the included instructions. A Baidu Cloud link is also provided (password: as84).

3.1 Three‑line code for descriptive statistics report

import report as rpt
# data encoding and import
# 300_300_0.xls is Wenjuanxing text data, 300_300_2.xls is Wenjuanxing numeric data.
data, code = rpt.wenjuanxing(['300_300_0.xls', '300_300_2.xls'])
# generate descriptive statistics report
rpt.summary_chart(data, code, filename=u'调研报告初稿')

The above code creates two files in the .\\out\\ folder: 调研报告初稿.pptx: descriptive statistics for each question, supporting single‑choice, multiple‑choice, ranking, matrix single‑choice, etc. 调研报告初稿.xlsx: statistical data (frequency and proportion) for each question.

3.2 Four‑line code for cross‑analysis report

import report as rpt
data, code = rpt.wenjuanxing()
save_dstyle = ['FE', 'TGI', 'CHI']  # choose metrics to save
rpt.cross_chart(data, code, cross_class='Q1', filename=u'性别差异分析', save_dstyle=save_dstyle)

This generates five files in .\\out\\: 性别差异分析.pptx: gender‑based differences for each question.

性别差异分析_百分比.xlsx

性别差异分析_FE.xlsx

性别差异分析_TGI.xlsx

性别差异分析_CHI.xlsx

3.3 Other useful functions

import report as rpt
# File I/O
data = rpt.read_data(filename)
code = rpt.read_code(filename)
rpt.save_data(data, filename, code)
rpt.save_code(code, filename)

# Single‑variable frequency table
t, t1 = rpt.qtable(data, code, 'Q1')
# Two‑variable cross table
t, t1 = rpt.qtable(data, code, 'Q1', 'Q2')

# Contingency table analysis
cdata = rpt.contingency(fo)

# Goodness‑of‑fit test
rpt.gof_test(fo, fe)

# Chi‑square test
rpt.chi2_test(fo, fe)

# Binomial confidence interval
rpt.binomial_interval(p, n)

# Automatic descriptive report
rpt.summary_chart(data, code, filename=u'描述统计报告', summary_qlist=None, max_column_chart=20)

# Automatic cross report
rpt.cross_chart(data, code, cross_class='Q1', filename=u'交叉分析', cross_qlist=None, plt_dstyle='TGI', save_dstyle=None)

Author: JSong, 2017‑02‑28