Research Impact Monitoring Tool
Creating research impact is increasingly important for Australian research as it ensures that
academic studies make a demonstrable contribution to the economy, society, culture,
public policy or services, health or the environment
(see: What is Research Impact).
Therefore, we need to develop tools that monitor our research impact and its translation within society. A proxy for research impact is certainly the amount to which research topics or key terms (e.g., "transformative work design", "cancer therapy", "Helicobacter pylori") or researchers that are associated with this type of research (e.g., "Sharon Parker", "Barry Marshall") are mentioned in the public media.
Given the exponential amount of information that appears regularly in the media and in social networks, this monitoring task is likely to create an information overload and unreasonable work demands if carried out by the researchers themselves. Therefore, a computational solution that condenses media-related information when tracking impact of research related activities is needed.
This computing project follows this logic and seeks to develop a tool that is able to continuously assess research impact by monitoring the degree to which research topics and key terms are mentioned in the media (Australian newspaper, websites) or in social networks (e.g., twitter or facebook).
A tool that continuously monitors (e.g., on a daily, weekly, monthly, yearly level) will allow researchers to track when and how their research impacts larger society. The challenge of this computing tool is therefore to transform large amounts of information (big data challenge) into a digestible format.
This computing project is divided in 3 steps.
-
Extraction of textual information from public sources
In the first step, the group will (in coordination with the contact person) select a set of text-based data sources (e.g., social networks, Australian newspapers, Wikipedia, regular company announcements like shareholder letters) that fulfil the following criteria:
-
The sources potentially provide time-sensitive information about research-related impact (i.e., daily, weekly, or yearly variations in new textual input)
-
The information is provided in text
The computing group will develop a program (e.g., crawler) that extracts the textual information in a systematic way (e.g., tab-formulated excel sheet). This data file also provides a time-stamp and supplemental information about the data source (e.g., web address, title).
-
Big data reduction: Transformation of textual information into a numerical index
In the second step, the systematically saved textual information (original data source) will be linked with a content-based lexical program (e.g., Linguistic Word Count and Linguistic Inquirer) in order to classify the number of research-related key-words that appear within the data-source and transform it into a numerical index (e.g., percentages of "key words" mentioned). The contact person will provide assistance with respect to the processes and rationale that are involved with the of text-to-number transformation. This assistance involves the provision of internal lexica that capture research impact key words for a specific area.
-
Monitoring of research-impact index
Finally, the group develops a graphical display that continuously monitors the index of research impact.
Ideally, the software solution will be able to link the three steps and automate the research impact tracking tool (e.g.: the researcher just selects the data source (e.g., "the Australian", step 1), types in key words or loads a key-word lexicon (i.e., definition of internal lexicon, step 2), selects the temporal resolution (i.e., days/weeks/months) and then receives the "monitoring-output").
The software package:
-
captures textual data from the various public sources (raw data);
-
Is able to extract "key words" in the raw (textual) data by means of a lexicon-based approach;
-
Generate a continuous monitoring (report) that illustrate the amount of "research terms" in those public sources
-
Should allow the user to trace back the original "raw data source", that is, changes in the impact-score only reflect a numerical change but the user may want to know in which context the research was mentioned (mixed methods approach)
-
Should be intuitive and user-friendly.
Client
Contact Person: Dr. Florian Klonek
Telephone: +61 8 6488 4585
Email: [email protected]
Preferred method of contact: E-mail or phone
Location: UWA Business School, Centre for Transformative Work Design, Room G39
Client Unavailability
None
IP Exploitation Model
The client wishes to use a Creative Commons CC BY-NC model to deal with IP embodied in the project.