REQUIREMENTS

Project backgroud and client

Sage publications is an American publishing company. It was founded in 1965 by Sara Miller McCune. Sage is a leading international provider of innovation, high quality content publishing more than 900 journals and over 800 new books each year across various subjects areas.

Project goals

Scientist write research paper which are then read by editors or publishers. These are then published depending on the question - will it be well received when it comes out and is it even worthy to be published. Editors have to peer review and go through all the papers manually which is time consuming and might get messy. Therefore, our client wants us to build a program (website) which makes this decision of publishing the new paper or not easier and faster (automated). The way we are approaching problem is by predicting a citation count using some of the data from the article. The citation count of a paper is the number of papers that cite it as a source. Citations are often used as currency , if you cite more , you get cited more and your paper is more likely to be seen.

Requirements gathering

We got an email and a brief about what our project was. But the brief was not very clear, therefore we decided to have an initial call with our client. We introduced ourselves and understood the root cause of the issue and what is the basic requirement for our project. We had a few follow up calls to get all the requirements and discuss our approach and structure for the problem. For our follow up meeting we drafted up some questions to get more pointers on the requirements.

    For example-
  • Are there any specific features u wany to have in the UI?
  • Who will be using this application. Is it only a sage publications application or would it be open to public ?
  • Where are we going to get the data set from? Etc

Persona

Use cases

There are two user cases

  1. Get citation prediction with abstract and title of the paper
    1. Upload file and enter title , choose the model
    2. Click on upload button
    3. Redirects to analysis page with the title , citation score and abstract
  2. Search for an existing research paper using DOI
    1. Enter the DOI in the search bar
    2. Click on the search bar
    3. Redirected to the search page - if a valid DOI is entered the details of the paper are shown on the page. Otherwise 'invalid DOI' is printed on the page


    3. The user case diagram is as follows

MoSCoW requirement list

Functional requirements



Id Requirements Priority
1 Output a number which is a citation count Must
2 Allow the user to upload a file Should
3 Allow the user to choose a file or drag -drop file on the web app Should
4 A search bar to search papers using DOI Should
5 Search using the author's name Could
6 Display H-index and author name with citation prediction if more information is provided Could


Non-Functional requirements



Id Requirements Priority
1 Usability: Clear and easy to navigate UI Must
2 Accuracy: The R2 for the prediction model should be greater than 0.3 Must
3 Use python as the backend language Must
4 Security: Add a login page Could