IT Innovation Contest

A team-based contest for creative IT solutions

the Paper Link for PubMed

Proposal Status: 

__Search and Get PDF Link Now__

 

This is a browser extension, a web app, also a mobile app!

-- built based on eUtils and related APIs of various web sites; Fast, Light & Mobile Friendly

 

#### Targeted User ####

The project targets to anyone frequently uses PubMed. It presents all the information to the users in one place, in real time, to help the users to save their valuable time.

 

#### Aims ####

* Direct access the PDF link of the related articles, on PubMed pages and any webpage with DOI/PMID/PMCID

* Integrate information for easy access and sharing

    ** When browsing on PubMed, display the impact factor, F1000 link, citation statistics, and any related information next to the article
    ** When browsing on PubMed, a button will be placed next to the article for copy and paste the article summary between applications
    ** When browsing on PubMed, provide links to email the PDF files or save the select articles to the Cloud services, including sharing to social networks if the user wants to

* Rich search experience when access from any mobile device, including smartphones and tablets, with single click to the PDF files, related articles, author publication statistics, etc.

* Alert service to notify users when new entry in PubMed has keyword matched their pre-entered list (alert will be delivered in real time; via email, Twitter direct message, Facebook wall post, etc.)

 

#### Current Progress ####

* Backends running on Google AppEngine and Amzon EC2, with >2,800,000 articles cached for quick response

* Browser extensions written by Javascript, HTML and CSS; open sourced on github

    ** Test version for Chrome, >2000 active users
    ** Test version for Firefox, >100 active users
    ** Test version for Safari, no statistics

* Testing the web app and the mobile app, running at http://www.thepaperlink.com

* Testing the Facebook App, running at https://apps.facebook.com/thepaperlink/

* Testing an API for calculating an author’s publication statistics; a page uses this API running at http://2.pl4.me/static/peaks-rev.html

* Keyword match based alert service tested with several users

 

#### To-do List ####

* Decrease the server latency, aiming for 100 milliseconds response time

* Improve the user interface of the web app, using Twitter Bootstrap

* Integrate support for Google Drive, and other widely adopted cloud storage services (currently when the user click “save it”, the file can be saved to Dropbox, but not to Google Drive, etc)

* Improve keyword based text matching and alert service, including customizable alert frequency, directly store keywords for alert if the user chose to, parallel processing capability on large user population

* Build a smarter alert service using text mining and machine learning; a prototype using Google Prediction API has been tested and proven to be promising - we will modify it to decrease the cost on large user population

* Improve the Javascript client http://pl4.me/js/ for browsers that do not have native extension/add-on support

* Improve the integration between Google Scholar and PubMed, such as sorting by citation count, sorting by journal impact, etc.

* Improve the integration with http://www.pubmeder.com to allow full text search in a user’s own article collection

* Add support to display the Paper Link information bar into other webpages, such as F1000 comments, Cochrane summaries, etc

* If time permits, build a native iOS app to decrease the network traffic required to use the mobile app

 

#### Team Member ####

* Developer: liang.cai@ucsf
* Designer: hao.wu@ucsf
* Tester: xiao.peng@ucsf

 

#### Questions Answered ####

* Why someone would be interested in using the Paper Link over PubGet?

    ** PubGet wants the users to browse the articles in their website, as a mean for revenue. The Paper Link provides many ways to access the information, without the need to change the existing reading habits.
    ** PubGet does have a mobile app, but it has much less function than the Paper Link.
    ** The Paper Link has better algorithm to extract the PDF links, which is the core of both services.
    ** The Paper Link has better integration with the social networks. Sharing the knowledge openly!
    ** There is no way to email the PDF file to yourself in PubGet.
    ** PubGet requires an institutional affiliation to check the PDF link, and requires visiting the site from a specific IP range to see the PDF file. The Paper Link does not need that. The Paper Link: displays the open access link to anyone from anywhere, allows the using of EZProxy, gives the PDF link to the user directly.
    ** The browser extension and addon of the Paper Link are open sourced. API of the Paper Link is free for personal use.
    ** The Paper Link was inspired by PubGet. It is built on the believe of Free, Open and Sharing.

* My NCBI has email alert service, why is the Paper Link trying to “reinvent the wheel”?

    ** My NCBI’s service is not in real time - an article might take days, even weeks to get to inbox
    ** My NCBI’s service is based on keyword matching, not by artificial intelligence
    ** My NCBI’s service only delivers the alert via email

Comments

Very nice time-saving utility that allows users to quickly retrieve relevant articles. The browser extensions are key since they allow functionality from within Pubmed itself. However, this is a team based contest and I do not see the listing of the team members for this project.

Just add the people behind the project. Any proficient iOS or Android developer are welcome to join.

Agree with prior commentor re: discussion on team. Would also be good to understand the targeted use cases and estimated impact. But sounds great - sounds like a worthy discussion with the CTSI team that brings UCSF Profiles to campus as well.

The project targets to anyone frequently uses PubMed - I think all the researchers in UCSF do. It presents all the information researchers interested in one place, and helps the users saving their valuable time.

It might be good to add a sentence or two to explain why someone would be interested in using Paper Link over PubGet. Not sure if the apparent cost recovery via Google Ads on the site is an issue for this contest. I suspect it's not a problem, but thought it worth mentioning in case I'm wrong. If you're looking for others to help, I might be interested in doing some of the JavaScript work.

* PubGet wants the users to browse the papers in their website, as a mean for revenue. The Paper Link proves many ways to access the information, without the need to change the existing habits.
* PubGet does have a mobile app, but it has much less function than the Paper Link.
* The Paper Link has better algorithm to extract the PDF links.
* The Paper Link has better integration with the social networks.
* There is no way to email the PDF file to yourself in PubGet.
* PubGet requires an institutional affiliation to check the PDF link, and requires visiting the site from a specific IP range to see the PDF file. The Paper Link does not need that. The Paper Link: present open access link to the user which can be read by anyone from any where, allow the using of EZProxy, give the link to the user directly.
* The browser extension and addon of the Paper Link are open source. API of the Paper Link is open to everyone.
* As for the Google Ads, it only shows on the web site when visiting from a Desktop. Mobile app and browser extension are free of Ads. It is for the sustainability of the service, because we are using Google App Engine and Amazon EC2 to store data and handle requests.

ONE MORE THING: at least right now, PubGet has not released a tool to record the articles a user read. The Paper Link utilizes an API from http://www.pubmeder.com for archiving, and under the permit of the user, will use that information to predict whether a new article will be interested by this specific user.

A very useful APP, which allows to get all the information at a glance. Cool!

Commenting is closed.