HackMD
Instructions for the Wiki Search Engine Task
The objective of this task is to integrate the SciBeh c19vax wiki with the Scibeh knowledge base.
Information about the tools
How are we linking up the two?
Each Wiki page has a specific focus. For example, the side effects page details the evolving situation around potential side effects associated with the COVID-19 vaccines, and the cultural differences page explores how and why attitudes towards and uptake of the vaccines may vary across cultures.
Our pages are fed by sources emailed to c19vaxwiki@gmail.com, which also go to the Knowledge Base. We have volunteers annotating the Knowledge Base and updating the pages. However, it would be good to also provide the public with a direct link from the Wiki to the Knowledge Base resources. To do so, we need to identify the best search terms that will generate the most relevant Knowledge Base items for each Wiki page.
Our Knowledge Base is also intended to add value compared to Google/Google Scholar by providing a more insightful and eclectic selection of items. We thus need to compare against Google/Google Scholar to assess whether we are achieving this aim.
What will I do as a volunteer?
Finding the right search terms can be a process of trial and error. So we need to try many different searches and assess the results. We want to see if they are actually relevant.
Right now we are in the process of setting up the search, so we will use the process below to try out different terms. We will adapt them with your feedback (thank you to those of you who have given us more insight on this!)
Search Process
Each time you try a search, these are the steps:
- Go to the Wiki page you signed up for.
- Read the page and identify the main issues so you can think of a set of search terms to try.
You might want to check the records spreadsheet before you do this to make sure someone else hasn't already tried your search term.
- Go to the knowledge base and type in your identified search term.
- Record the exact search terms you used in the spreadsheet.
- Also write down the number of annotations that appeared for the search.
- Fill out the 'ratings' columns of the spreadsheet whether you think the search gives enough relevant results, too many irrelevant results etc.
- Add any other notes you have about this search in the 'notes' column of the spreadsheet.
- Repeat this process with different search terms.
Example
In this example, I am going to try the search terms side effects to search the Knowledge Base for items relevant to the Wiki COVID-19 side effects page. [Steps 1 & 2]
[Step 3] I typed my search terms into the knowledge base:
Here you can see it turned up 52 matching annotations. A quick scan shows it has many relevant items but also some that are not related to COVID-19.
[Step 4-5] I recorded the search link in the spreadsheet and the number of annotations that turned up.
[Step 7a] I rated the relevance/irrelevance of the items from the knowledge base. (I felt it had a middling number of relevant items, but quite a lot of irrelevant ones–-don't worry too much about this, we know it will be subjective, so just rate however you think!)
Editing the knowledge base/tags
When looking up search terms for the Wiki use case, think about what you are actually trying to find when exploring the knowledge base. Are you missing certain entries that for e.g., you know you annotated? When exploring the different search terms also have a look at the different annotations and feel free to add tags that you feel like are missing from a specific annotation.
Volunteers on this team
- Jacqueline Krauss
- Cheyenne Chooi
- Chaeyeon Lim
- Zoe Ikeotuonye
- Natasja Derby-McCabe
- Luning Bai
- Sophia Sterckx
- Paul McIlhiney
- Emily Biggs
- Li Qian Tay
- Will Hughes
