This presentation was given by guest lecturer Pavel Kasyanov of Clarivate Analytics during the fifth session of the NISO Spring training series "Working with Scholarly APIs." Session Five, Web of Science, was moderated by Phill Jones of MoreBrains Cooperative and held on May 26, 2022.
3. 3
The same article record
• This is only a fragment of the full
record data
• Every piece of metadata is stored
in a dedicated field and can be
retrieved for further data
operations
But the metadata obtained using Web of Science Expanded API
4. Web of Science APIs: Primary Use Cases
IR and Public
Researcher
Engagement
Platforms
3rd party integration –
FAR/CRIS/RIMS and
Discovery Services
Publication & Research
Presentations / Findings
Measuring and
Modeling the
Dynamics of Science
Hypothesis testing and
validation
Combining scholarly
impact data with
other external
datasets
Citation Networks &
Mapping
Research Intelligence &
Strategy / Planning
Scientific Trends and
Impacts over time
Data Integration (on-going) Research Projects (one-off)
Dashboards & Visuals
4
5. 5
Difference between user interface and API
API (Application Programming Interface)
Any task on which computer works more efficiently:
• Routine calculations
• Extracting only necessary data fields
• Combining Web of Science data with external data
for further analysis
Web of Science platform user interface
Any task designed for human brain:
• Running a topical search in Web of Science
• Filtering results
• Selecting the most relevant search results
• Clicking through to the full text document to read it
General approach
10. 10
Insert footer
How to get your API key
• Register (your Web of
Science credentials
should work)
• Create an application
• Select an API and click
“subscribe”
developer.clarivate.com
13. Data Integration is an important part
of API use cases – but API and raw
data are more than just that
With API or raw data you can expand
your analytics far beyond the
capabilities of the user interface
13
14. 14
Simplest example
Processing Web of Science Core Collection data the way you want
Web of Science data
A simple self-made or
community-made program or
algorithm
Answers to your
unique and specific
questions
16. Example #1:
“Which author IDs are created for the
researchers in my organization? How
up-to-date are they?"
16
17. 17
Simple code for extracting the necessary data
• The user simply enters their organization profile
name into the code
• And runs the program
• The code is freely available at this GitHub link – you
only need Web of Science Expanded API key to run it
18. 18
Result:
Including their ResearcherID
A .csv table containing all the author records for our organization
Their ORCID
A link to their Web of Science
Author Record
Their list of Web of Science
documents
And whether their Web of
Science author record is
claimed or not
19. Example #2:
“Can we calculate the H-index for our
researchers - but excluding their self-
citations?”
19
20. 20
Simple code identifying self-citations
• The user simply enters the author’s name,
ResearcherID, or ORCID
• And runs the code
• The code is freely available at this GitHub link – you
only need Web of Science Expanded API key to run it
And removing them from the analysis
21. 21
Result:
• This gets printed out
by the program:
• And this is the .csv file:
• This is where the difference comes from:
Summary printed out by the program and details saved to a .csv file
22. Example #3:
“We want to calculate our
organization’s research output using
fractional, not whole, document
counting”
22
23. 23
Simple code for fractional counting
• The user simply enters their organization profile name
and publication years
• And runs the code
• The code is freely available at this GitHub link – you
only need Web of Science Expanded API key to run it
At the organizational level
24. 24
Result: Bonus: 2 additional
meaningful metrics
Also a .csv file – with individual documents data and a possibility to unite them
into a pivot table
25. 25
Other examples
• City-level collaboration analysis
– Which cities is a certain research topic
concentrated in?
– Which cities does a specific organization
collaborate with?
– In which cities does a specific organization
concentrate its research?
• Keyword analysis
– Which keywords appear on the most recent
research on a given topic? (Which research trends
are emerging in a specific research area?)
• etc.
By processing Web of
Science data, we can:
• Create our own
derivative metrics
• Quickly perform
calculations that would
be too time-consuming
if done manually
For processing Web of Science
data obtained via API
26. Almost any analysis with Web of
Science data that you can think of –
even if it is not yet implemented in
Web of Science user interface – is
possible using API or Raw Data
26