We believe that collaboration is the key factor for impactful high-quality research. CollabReSearch is a platform to support you in the hard task of finding potential contributors to leverage your research and career!

What it does

Collab Research is a search engine to help you find potential contributors to leverage your research. The results are shown in a georeferenced map with 3 dimentions:

  • Location - This can be seen in the map presented by the platform
  • Relevance of the researcher publications in the area matched by the query string - This is represented by the color of the circle representing the researcher (Green > Yellow > Orange > Red)
  • Productivity and citation impact of the researcher - Represented by the size of the circle representing the researcher (the bigger the circle, the higher his/her hindex)

The user can click on the circles representing the researchers to get more information about them, like

  • hindex value
  • Affiliation
  • Total number of citations
  • Areas of interest
  • Links to Scopus and Google Scholar profiles
  • Google Scholar profile picture

How we built it

Since we did not have any access to CNPq's Lattes data (unfortunately the date is not open for machines and require captchas to be accessed) we needed to use 3 different data sources to retrieve all the data needed to complete this prototype:

First, we retrieve articles matching the queries entered by the user (relevance provided by Mendeley). Then we retrieve all the authors of those articles and give them points based on the number of views for the matching papers, also provided by Mendeley. Finally, we query Scopus (API) and Google Scholar (scrapper python lib) to retrieve more information about each researcher.

Since Scopus API is important to our process and it is quite slow, unstable, and expensive, we decided to display at most 10 researchers in the map.

We used the Python programming language to fetch our data and provide a back-end to our application using the Flask framework.

We used Google Maps API to plot our data and the Material Design Light library to make our front-end beeeautiful!!!!

Challenges we ran into

Mendeley API does not always provide reliable information, so we needed to cross information with the other two platforms, where researchers had different IDs and we needed to use some heuristics in to determine parts of our data.

Our designer learned JS, GIMP and CSS in 10 hours! \o/ (we are all pretty much backend developers)

Accomplishments that we're proud of

  • The interface is quite beautiful for a bunch of low level programmers!
  • We could finish the hackathon with a nice MVP!

What we learned

  • JavaScript :)
  • RedBull keeps you ON!

What's next for Collab Research

We would love to be able to use CNPq data :(

We need to improve how we rank researchers!

Try it!

Note that this is running in a debugging mode, so the requests may take longer than you would like to wait.

Source Code and License

Collab Research is licensed under the General Public License (GPLv3) and its source code is available in GitHub

Built With

Share this project: