Updated: Feb 8, 2021
This is the final part of a 3 parts blogpost series in which we show you how our software works in more details.
In the first part, we covered how to log in to the HubScience system, how you can create a project, invite members, set their roles, what sources you can use to upload documents, and how to upload them, and what you can see on the Documents panel.
In the second part, we showed you how the Annotation tool panel – the soul of our product- and the Dictionary panel works.
In this post, you will get to know more about the final two panels, Knowledge and Preannotation rules.
The Knowledge panel is full of tools that can help you analyse the automatically and manually annotated documents’ results.
This page summarises the expressions annotated within the project. First, you can see them in a list, and next to the words, you can check how many of them were found.
Once you click on any of the words, you can see more info, such as which documents included the words. Clicking on one of them, you will see them within the original text.
The order can be changed by alphabetic order or how many times it has been found, and you can also see which category and sub-category it’s been added to.
It’s also possible to screen the shown elements, you can view the automatic or manual annotations, screen by articles, timeframe or users, meaning who annotated what in the project.
The easiest way to do it is using the circle diagram on the right, clicking on each and every part of it will show you different results.
In advanced screening mode you can combine any analysing criteria with any other criteria:
There are further possibilities to search within your results, and you can see in the spreadsheet what details did you screen by.
You can export the whole-, or screened spreadsheet too, as a CSV file compatible with Excel.
Another visualisation option is viewing the results in graphs where you can see how the articles share your annotated keywords. We suggest you use it under 2000 elements shown to make it easier to analyse.
There are built in screening options, one is searching for keywords, expressions and numeric data that were in all the documents, another option is to view the ones that were in more than one documents.
HubScience can automatically recognise and collect over 2000 units and the connected data, and show them in the knowledge graph. The knowledge graph has built in screening rules with what it easily helps you find common points in documents or find unique results.
Intersections – Connected to exactly one document:
The graph view is really practical when you need to get information from many publications and find relevant articles that are the most useful. It saves you a lot of time.
It’s also connected to Wiki, so if there’s a link and you click on it, you can see the other relevant info connected to our topic, and it can open other sources. Information from Wiki database are shown in elliptic nodes, so you can clearly see that it isn’t from any of the project documents.
This helps you find more info, even those that you didn’t think of when starting your project.
On the right side of the graph you can see the information panel that includes the highlighted Wiki article and further searching suggestions. These are active links so once you click on them it jumps to the whole article in a new window, supporting you in getting info faster and more effectively.
There’s also a search bar (‘Query by occurences’ button) in the Knowledge panel. You can see suggestions after annotating for your keywords, and connect them together with ‘AND’ and ‘OR’ orders.
You can save your search criterias and run them over and over again anytime:
The list of results will show which documents are connected and which other annotations are relevant with them.
You can also see these results in tables and export them as image or data, and you can print it too.
Analyse your data in spreadsheets, diagrams and graphs:
When you click on one element it opens up for you.
HubScience is unique in serving you the possibility of building patterns specific to projects, with what you can create a connection net between parts of the texts that you cannot process in any other way.
There are built in rules that you can extend with your own rules. You can define a keyword’s role towards another keyword in a sentence, for example in front or behind.
You can add these to categories, so you don’t need to do them one by one, and so the rules will be given to all the words in that category.
This is how we can make sure that text mining with our software gives you high quality results.
These are the basic and most important parts our software knows. Thank you for joining us in these introduction posts.
Should you have any questions, or want a demo for your company, don’t hesitate to contact us at firstname.lastname@example.org!