| author | title | |
|---|---|---|
|
`litreviewer`: A Python Package for Review of Literature (RoL) |
Secondary research involves the summary, collation and/or synthesis of existing research. Secondary research is contrasted with primary research in that primary research involves the generation of data, whereas secondary research uses primary research sources as a source of data for analysis. A notable marker of primary research is the inclusion of a "methods" section, where the authors describe how the data was generated. Common examples of secondary research include textbooks, encyclopedias, news articles, review articles, and meta analysis. [@Ormsby] [@Andrews] When conducting secondary research, authors may draw data from published academic papers, government documents, statistical databases, and historical records.
A Literature Review is a systematic and comprehensive analysis of books, scholarly articles, and other sources relevant to a specific topic providing a base of knowledge on a topic. A literature review is an overview of the previously published works on a specific topic. Literature reviews are designed to identify and critique the existing literature on a topic to justify your research by exposing gaps in current research. This investigation should provide a description, summary, and critical evaluation of works related to the research problem and should also add to the overall knowledge of the topic as well as demonstrating how your research will fit within a larger field of study. A literature review should offer a critical analysis of the current research on a topic and that analysis should direct your research objective. This should not be confused with a book review or an annotated bibliography both research tools but very different in purpose and scope. A Literature Review can be a stand-alone element or part of a larger end product, know your assignment. The key to a good Literature Review is to document your process.
Mostly this phrase may refer to a full scholarly paper or a section of a scholarly work such as a book, or an article. Either way, a literature review is supposed to provide the researcher/author and the audiences with a general image of the existing knowledge on the topic under question. A good literature review can ensure that a proper research question has been asked and a proper theoretical framework and/or research methodology have been chosen. In other words, a literature review serves to situate the current study within the body of the relevant literature and to provide context for the reader. Literature review is often a part of scholarly work, including in the preparation of a thesis, dissertation, or a journal article. Literature reviews are also common in a research proposal or prospectus. [@Baglione]
Literature reviews are secondary sources and do not report new or original experimental work. Most often associated with academic-oriented literature, such reviews are found in academic journals and are not to be confused with book reviews, which may also appear in the same publication. Literature reviews are a basis for research in nearly every academic field.
The main types of literature reviews are: evaluative, exploratory, instrumental, systematic review and . [@Adams]
Systematic reviews are a type of review that uses repeatable analytical methods to collect secondary data and analyze it. Systematic reviews are a type of evidence synthesis which formulate research questions that are broad or narrow in scope, and identify and synthesize data that directly relate to the systematic review question. Systematic reviews critically appraise research studies, and synthesize findings qualitatively or quantitatively.[@Armstrong] Systematic reviews are often designed to provide an exhaustive summary of current evidence relevant to a research question.
The process of creating a systematic review can be broken down into between four and 15 tasks depending on resolution. [@Cochrane] Systematic reviews are conducted through a robust but slow and resource-intensive process. The result is that undertaking a systematic review may require a large amount of resources and can take years. Proposed decision support systems for systematic reviewers include ones that help the basic tasks of systematic reviews. The full automation of systematic reviews that will deliver the best evidence at the right time to the point of care is the logical next step. Indeed, research into automatic systematic review systems is distinct from - and related to - research on systematic reviews. [@Tsafnat]
A meta-analysis is a statistical analysis that combines the results of multiple scientific studies. Meta-analysis can be performed when there are multiple scientific studies addressing the same question, with each individual study reporting measurements that are expected to have some degree of error. The aim then is to use approaches from statistics to derive a pooled estimate closest to the unknown common truth based on how this error is perceived.
Existing methods for meta-analysis yield a weighted average from the results of the individual studies, and what differs is the manner in which these weights are allocated and also the manner in which the uncertainty is computed around the point estimate thus generated. In addition to providing an estimate of the unknown common truth, meta-analysis has the capacity to contrast results from different studies and identify patterns among study results, sources of disagreement among those results, or other interesting relationships that may come to light in the context of multiple studies. [@Greenland]
Text mining, also referred to as text data mining, similar to text analytics, is the process of deriving high-quality information from text. It involves "the discovery by computer of new, previously unknown information, by automatically extracting information from different written resources." Written resources may include websites, books, emails, reviews, and articles. High-quality information is typically obtained by devising patterns and trends by means such as statistical pattern learning. According to Hotho et al. (2005) we can differ three different perspectives of text mining: information extraction, data mining, and a KDD (Knowledge Discovery in Databases) process. [@Hotho] Text mining usually involves the process of structuring the input text (usually parsing, along with the addition of some derived linguistic features and the removal of others, and subsequent insertion into a database), deriving patterns within the structured data, and finally evaluation and interpretation of the output. Typical text mining tasks include text categorization, text clustering, concept/entity extraction, production of granular taxonomies, sentiment analysis, document symmetrization, and entity relation modeling (i.e., learning relations between named entities).
Text analysis involves information retrieval, lexical analysis to study word frequency distributions, pattern recognition, tagging/annotation, information extraction, data mining techniques including link and association analysis, visualization, and predictive analytics. The overarching goal is, essentially, to turn text into data for analysis, via application of natural language processing (NLP), different types of algorithms and analytical methods. An important phase of this process is the interpretation of the gathered information.
A typical application is to scan a set of documents written in a natural language and either model the document set for predictive classification purposes or populate a database or search index with the information extracted. The document is the basic element while starting with text mining. Here, we define a document as a unit of textual data, which normally exists in many types of collections. [@Feldman]
A Web crawler, sometimes called a spider or spider-bot and often shortened to crawler, is an Internet bot that systematically browses the World Wide Web, typically operated by search engines for the purpose of Web indexing (web spidering). Web search engines and some other websites use Web crawling or spidering software to update their web content or indices of other sites' web content. Web crawlers copy pages for processing by a search engine, which indexes the downloaded pages so that users can search more efficiently. Crawlers consume resources on visited systems and often visit sites without approval. The number of Internet pages is extremely large; even the largest crawlers fall short of making a complete index. For this reason, search engines struggled to give relevant search results in the early years of the World Wide Web, before 2000. Today, relevant results are given almost instantly. Crawlers can validate hyperlinks and HTML code. They can also be used for web scraping and data-driven programming.
A Web crawler starts with a list of URLs to visit, called the seeds. As the crawler visits these URLs, it identifies all the hyperlinks in the pages and adds them to the list of URLs to visit, called the crawl frontier. URLs from the frontier are recursively visited according to a set of policies. If the crawler is performing archiving of websites (or web archiving), it copies and saves the information as it goes. The archives are usually stored in such a way they can be viewed, read and navigated as if they were on the live web, but are preserved as 'snapshots'.
The number of possible URLs crawled being generated by server-side software has also made it difficult for web crawlers to avoid retrieving duplicate content. Endless combinations of HTTP GET (URL-based) parameters exist, of which only a small selection will actually return unique content.
The behavior of a Web crawler is the outcome of a combination of policies: (1) a selection policy which states the pages to download, (2) a re-visit policy which states when to check for changes to the pages, (3) a politeness policy that states how to avoid overloading Web sites, (4) a palatalization policy that states how to coordinate distributed web crawlers.
Web scraping, web harvesting, or web data extraction is data scraping used for extracting data from websites. The web scraping software may directly access the World Wide Web using the Hypertext Transfer Protocol or a web browser. While web scraping can be done manually by a software user, the term typically refers to automated processes implemented using a bot or web crawler. It is a form of copying in which specific data is gathered and copied from the web, typically into a central local database or spreadsheet, for later retrieval or analysis.
Web scraping a web page involves fetching it and extracting from it. Fetching is the downloading of a page (which a browser does when a user views a page). Therefore, web crawling is a main component of web scraping, to fetch pages for later processing. Once fetched, then extraction can take place. The content of a page may be parsed, searched, reformatted, its data copied into a spreadsheet or loaded into a database. Web scrapers typically take something out of a page, to make use of it for another purpose somewhere else. An example would be to find and copy names and telephone numbers, or companies and their URLs, or e-mail addresses to a list (contact scraping).
Traditional approaches to reviewing literature may be susceptible to bias and result in incorrect decisions. Systematic reviews have been introduced as a more rigorous approach to synthesizing evidence across studies; they rely on a suite of evidence-based methods aimed at maximizing rigor and minimizing susceptibility to bias. Despite the increasing popularity of systematic reviews the evidence synthesis methods continue to be poorly applied in practice, resulting in the publication of syntheses that are highly susceptible to bias. [@Haddaway]
Following are the common problems in RoL.
-
Too descriptive
-
Lack of quality sources
-
A by-article structure
-
No link to research question
-
Un-grounded statements
litreviewer is a Python package which provides few methods to (1)
crawl (handling seeds), (2) collect (web frontier) and (3) archive
(snapshots). litreviewer is essentially a collection of few Python
modules as such this package is extensible; which means this package can
provide few text mining functions such as Tokenization, Stemming and
Lemmatization. Apart from this, the package provides few functions
related to statistical methods. Some of the essential modules are as
follows:
-
gscholar: a web crawler for google scholar. -
gsearch: a web crawler for google search engine. -
litutilities: a module for textual analysis and text conversions. -
literstat: a module for data analysis. -
cleandata: a module for NLP and text cleaning.
litreviewer has two web crawlers: (1) gscholar and (2) gsearch.
gscholar is a Python module which is useful to crawl across google
scholar and gsearch is another Python module which is useful to crawl
across google search. gscholar has a class gScholar which performs
both crawling and scraping tasks. This module possess certain
methods to perform crawling activity and obtain information from
scholarly sources. Few methods in this module are as follows:
-
makeQuery(): is a method that makes GET request for search query (known as seeding). A search query is any topic a researcher is intending to perform at google scholar. -
getLinks(): obtains links also called as web frontier. -
getTitles(): a utility function that scrape titles from various sources. -
openLink(): a utility function that helps opening a given link from the list of links. -
getText(): a utility function that helps extracting (through scraping) text from a give source (link).
The other module gsearch is useful to obtain (using crawling and
scraping) information from google search engine. This module has a
class known as gSearch which possess few methods which are as follows:
-
getUrl(): useful to to make a url for GET request (used for seeding). -
getTitlesAndLinks(): useful for obtaining (scraping) titles and links. -
getLinks(): obtains links for further queries from actual data obtained from scraping. -
openLink(): useful to open a given link from list of links. -
saveTextFromLinks(: saves data from a given web source using web scraping. -
downloadPDF(): useful for downloading PDF files from web. -
saveMultipleFiles(): useful to obtain text from multiple sources (links) using web scraping.
Apart from these modules there are other few modules which helps in performing certain data processing such as (1) obtaining data in text files (web scenarios), (2) concatenating data and forming a text corpus, (2) performing statistical analysis. These modules can be found in the package home directory.
This section explains as how to use these module together with their classes and methods. I am going to work upon present world crisis i.e., Covid-19 world pandemic.
WHO has been broaching both heard immunity and healthcare capacity
through time to time mandates. [@who1] [@who2] \cite{}{=latex}
Covid-19 pandemic can be mitigated using immunization and health
capacity. Let Covid-19 pandemic is a scenario and both immunization
and healthcare capacity are constructs. In which case it requires two
search queries; which possibly can be:
::: description "The impact of healthcare capacity on Covid-19 mitigation"
"The impact of immunization on Covid-19 mitigation" :::
As far as research methods are concerned; these data obtained from these search queries can be analyzed by using few statistical tests such as t test together with cluster analysis (using 2 cluster solution). The respective hypotheses for this analyzes can be as follows:
::: description "$H_{1}$: The healthcare capacity does not impact Covid-19 mitigation"
"$H_{2}$: The immunization does not impact Covid-19 mitigation"
"$H_{3}$: The difference between impacts of healthcare capacity and immunization is insignificant"
"$H_{4}$: Healthcare capacity and immunization can't explain Covid-19 mitigation" :::
Above study assumptions can be evaluated using few methods available
from litreviewer module using text mining. The very first activity is
to obtain data using web crawling and scraping. The following code
snippet can be useful for performing the aforementioned techniques.
[]{#listing1 label="listing1"}
import os
path = os.getcwd()
print(path)
title = 'impact of healthcare capacity on Covid-19 mitigation'
num_links = 10
duration = 1
gs = gSearch(query = title, num_links=num_links, duration=duration)
titles = gs.getTitlesAndLinks()['titles']
links = gs.getTitlesAndLinks()['links']
linksch = gs.getLinks(links)
The os base module is used to create a path variable. This variable
is very essential in order to save data as in text corpus. the title
is search query. num_links and duration are other parameters which
needs to be used for creating a GET request for web crawler. gs is
object of the class gSearch which is used for various other tasks
while obtaining links (web frontier) and titles (data). The object
linksch has the all those links that were scrapped from the web.
It is possible to know about GET query using print(gs.getUrl()). This
search query has returned with 13 links from data scraping (use
print(len(linksch)) to know about number links). We can obtain the
whole list of links (frontier) using following Python snippet.
for i in linksch:
print(linksch.index(i), i)
Finally the data can be saved as in text corpus using the following Python code snippet.
for i in range(13):
gs.saveTextFromLinks(linksch, path, link_num= i, name=f'{i}.txt')
The corpus can be found inside a folder with name output in the
package's home directory.
D:\Work\...> dir
Volume in drive D is Data
Volume Serial Number is ...
Directory of D:\Work\...>
29-05-2021 13:02 3,994 cleandata.py
30-05-2021 12:43 7,269 gscholar.py
30-05-2021 12:54 4,952 gsearch.py
30-05-2021 12:50 5,148 literstat.py
29-05-2021 16:08 5,737 litutilities.py
30-05-2021 12:49 <DIR> output
03-03-2021 15:23 0 __init__.py
30-05-2021 12:50 <DIR> __pycache__
10 File(s) 28,294 bytes
4 Dir(s) 700,159,221,760 bytes free
You can find a folder with name output which did not exist earlier. In
side this folder you can find various data files (in .txt format) which
has data which is scraped from all those links obtained earlier which
are available through the object linksch.
We are going to address study assumptions using few statistical tests and a cluster analysis. The data so obtained as in various text files (each file for a link) is actually a heap of text that needs to be collated as in a single file (also known as text corpus). This file now will be converted into data table. The data tables so obtained are known as term vectors or word frequency tables. The following code helps in (1) text concatenation, (2) analysis.
{python}
appendText(path)
wfreq = listOfWordsWithFreqs(path, print_words=True, nw=10)
datatbl = makeTables(wfreq, path, file_name = 'termmat')
from literstat import *
print(computeIQR(datatbl['freq']))
print(oneSampleTTest(datatbl['freq']))
clusanal = uniVarClusterAnalysis(datatbl, nc=1)
print(clusanal)
The output for the above code snippet is going to be as follows:
[('COVID-19', 447), ('Information', 16), ('Public', 40), ('health', 404), ('information', 58), ('(CDC)', 2), ('Research', 34), ('(NIH)', 2), ('SARS-CoV-2', 26), ('data', 162)]
83.0
Ttest_1sampResult(statistic=13.180014323782602, pvalue=2.5560862194761457e-39)
{'labels': array([0]), 'centroids': array([[447., 16., 40., ..., 1., 1., 1.]]), 'score': -0.0, 'silhouette_score': "Can't compute silhouette_score!"}
The very first line in the above code snippet is the sample term
vector. This is the output generated from the method
listOfWordsWithFreqs() for parameters print_words=True, nw=10. The
numeric output 83 is inter-quartile deviation. The third line has
results for T Test performed on data vector (word frequencies i.e.,
datatbl[’freq’]). The P Value is close to zero which means the sample
mean significantly differs from population mean (
The below code snippet retrieves the data from the web for search query, "impact of immunization on Covid-19 mitigation".
title = 'impact of immunization on Covid-19 mitigation'
num_links = 10
duration = 1
gs = gSearch(query = title, num_links=num_links, duration=duration)
titles = gs.getTitlesAndLinks()['titles']
links = gs.getTitlesAndLinks()['links']
linksch = gs.getLinks(links)
list_for_d = [0, 2, 3, 4, 5, 7]
for i in list_for_d:
gs.saveTextFromLinks(linksch, path, link_num= i, name=f'{i}.txt')
The above code snippet is similar to the one in the section
5.3.1{reference-type="ref" reference="healthcap"}. One
exception is that the files were retrieved using a list called
list_for_d. This list has indixes of links in the object linksch.
The missing numbers in this list denotes PDF files. PDF files were not
amenable for data analysis that is why they are not used in the
analysis. However, there is one particular method known as
downloadPDF(). This method is useful and can be used to scrape the
data manually. The results are as below:
[('Washington:', 1), ('A', 21), ('multi-institutional', 1), ('study', 21), ('suggests', 1), ('that', 81), ('prioritizing', 4), ('vaccine', 38), ('delivery', 4), ('to', 251)]
There was a problem while writing data to file. However you can safely neglect the same.
8.0
Ttest_1sampResult(statistic=11.573386984865788, pvalue=2.1354850040837868e-30)
{'labels': array([0]), 'centroids': array([[ 1., 21., 1., ..., 1., 1., 1.]]), 'score': -0.0, 'silhouette_score': "Can't compute silhouette_score!"}
The IQR is 8. T Test rejects null hypothesis that the
>>> df1 = pd.read_csv(filepath)
>>> df2 = pd.read_csv(filepath, encoding = "ISO-8859-1")
>>> from literstat import *
>>> twoSampleIndTTest(df1['freq'], df2['freq'])
Ttest_indResult(statistic=3.4744681151293126, pvalue=0.0005948270668645317)
>>> print(moodsTest(df1['freq'], df2['freq'], ties=False))
{'g': 49.03948368696504, 'p': 2.508615958325688e-12, 'med': 8.0, 'tbl': array([[98, 35],
[45, 98]], dtype=int64)}
>>> bartlettTest(df1['freq'], df2['freq'])
{'stat': 331.99366758595477, 'p-value': 3.5364522475713513e-74}
>>> df = pd.DataFrame({'v1':df1['freq'], 'v2':df2['freq']})
>>> biVariateClusterAnalysis(dfwona)
{'labels': array([1, 1, 2, 1, 2, 2, 1, 0, 1, 0, 0, 1, 0, 1, 2, 1, 1, 1, 2, 2, 1, 2,
1, 1, 0, 2, 2, 0, 0, 2]), 'centroids': array([[3.71428571, 1.71428571],
[8. , 4.46153846],
[4.1 , 6.8 ]]), 'score': -128.5879120879121, 'silhouette_score': 0.40949178182663026}
The above code snippet is related to bivariate analysis on two data
variables i.e., healthcare capacity and immunization. The T Test shows
evidence in support of alternative hypothesis i.e.,
{#clusterplot height="8cm" width="9cm"}
The study has successfully collected evidence in support of the hypotheses. From the univariate and bivariate analysis it is clear that the sample variables i.e., healthcare capacity and immunization has criterion validity and moreover there is sufficient evidence in support of the study hypotheses that both healthcare capacity and immunization do impact disease mitigation. All the test results are statistically significant. The 2 cluster solution from cluster analysis shows evidence that the data has information related to two distinct study constructs i.e., healthcare capacity and immunization.
The package litreviewer found to be effective while performing text
mining. The methods defined inside the modules found to work effectively
while using for web crawling and scraping. Modules associate with
statistical analysis and other utility functions related to text
concatenation and text corpuse found to have worked effectively. The
efficacy of the package was proved with the help of a case study related
to Covid-19. The package has found to have proved the validity of data
along with offering insightful outcomes from the same.
The package is still in development stage. Few methods or functions related to data cleansing need to be pruned well so as to minimize noise in the data sets. There are few methods or functions related to visualization. However, these functions supports very basic types of plots such as bar plot, pie plot, box plot and scatter diagrams. Whereas, data analysis is concerned, there is still scope for inducting methods associated with machine learning while performing bivariate or multivariate analysis. There is also scope for other qualitative analysis such as correspondence analysis, multidimensional analysis, principal component analysis etc.
::: thebibliography 100
Ormsby, Tim. "Search Smart: Types of Research: Secondary Research". flinders.libguides.com{.uri}.
Andrews, Camille. "LibGuides: Scholarly Literature Types: Primary vs. Secondary Articles". guides.library.cornell.edu.
Baglione, L. (2012). Writing a Research Paper in Political Science. Thousand Oaks, California: CQ Press.
Adams, John; Khan, Hafiz T A; Raeside, Robert (2007). Research methods for graduate business and social science students. New Delhi: SAGE Publications. p. 56. ISBN 9780761935896.
Armstrong R, Hall BJ, Doyle J, Waters E (March 2011). "Cochrane Update. 'Scoping the scope' of a cochrane review". Journal of Public Health. 33 (1): 147�50. doi:10.1093/pubmed/fdr015. PMID 21345890.
The Cochrane Collaboration: Cochrane Handbook for Systematic Reviews of Interventions. 2011,http://www.cochrane.org/training/cochrane-handbook, 51.
Tsafnat, G., Glasziou, P., Choong, M.K. et al. Systematic review automation technologies. Syst Rev 3, 74 (2014). https://doi.org/10.1186/2046-4053-3-74
Greenland S, O' Rourke K: Meta-Analysis. Page 652 in Modern Epidemiology, 3rd ed. Edited by Rothman KJ, Greenland S, Lash T. Lippincott Williams and Wilkins; 2008.
Hotho, A., Nürnberger, A. and Paaß, G. (2005). "A brief survey of text mining". In Ldv Forum, Vol. 20(1), p. 19-62
Feldman, R. and Sanger, J. (2007). The text mining handbook. Cambridge University Press. New York.
Haddaway, N.R., Bethel, A., Dicks, L.V. et al. Eight problems with literature reviews and how to fix them. Nat Ecol Evol 4, 1582�1589 (2020). https://doi.org/10.1038/s41559-020-01295-x
WHO. Immunization in the context of COVID-19 pandemic. Obtained from https://apps.who.int/iris/bitstream/handle/10665/331590/WHO-2019-nCoV-immunization_services-2020.1-eng.pdf.
WHO, Implementation of mass vaccination campaigns in the context of COVD-19. Obtained from https://www.who.int/docs/default-source/coronaviruse/framework-for-decision-making-implementation-of-mass-vaccination-campaigns-in-the-context-of-covid19-slide-deck.pdf?sfvrsn=438dccc8_2
Abideen, A. Z., et al. Mitigation strategies to fight the COVID-19 pandemic�present, future and beyond. Retrieved from https://www.emerald.com/insight/content/doi/10.1108/JHR-04-2020-0109/full/html :::