Leandro Contreras

January 23, 2023

Creating Centralized Platforms: ULisboa’s Scholar

In 2020, Técnico Lisboa, a faculty of the University of Lisbon/ULisboa, launched a repository to centralize the University's scientific output, uniting 50,000 users from 18 faculties in a single, modern platform.

Project Context

ULisboa, Portugal’s largest university group and one of Europe’s biggest, serves over 50,000 students from 100+ countries across 18 faculties. For years, its schools stored scientific work in separate repositories, creating fragmented access to information. ULisboa needed a unified platform to centralize research, allowing the community to store and import work from other databases for simplified access.

Company & Product

Técnico Lisboa, part of ULisboa's 18 schools, is Portugal's largest university for Architecture, Engineering, Science, and Technology, with a community of over 10,000. Its tech team developed Scholar, a repository for collecting, preserving, and sharing the community's scientific work. Using a self-archiving model, authors deposit their work, supported by tools for detecting duplicates and metadata issues, ensuring reliable access to knowledge.

View Product

Team

2 Designers, 3 Fullstack Engineers and 2 Backend Engineers. My role grew from support to Lead Product Designer, where I coordinated the tech team with the Engineering Manager, and spearheaded UX research.

Timeline

Continuous iterations in a waterfall model (from 2020 until 2022)

Impact

30155 Total Indexed Authors, 110059 Total Indexed Publications, +7% Publication Indexing Rate

The Challenge

A repository is a collection of scientific work, maintained by a single entity — a university or a research unit. But what if a university scales and contains 18 faculties, each with their own singular ways and platforms to store this information?

Imagine you’re an Architecture student at Técnico Lisboa, eager to explore eco-friendly solutions by connecting Biology and Architecture. You know that the Faculty of Sciences holds valuable resources, but to access their repository, you need a separate account on an unfamiliar system to navigate.

Now, multiply this by 18 faculties, each managing their own repository with different processes. This situation limits access, wastes time, and complicates research for ULisboa’s 50,000-strong community of students, researchers, and faculty.

So, the Técnico Lisboa tech team was tasked to centralize scientific production into a unified repository — Scholar — designed to streamline access and ease information sharing across disciplines.

Our Principles

When starting to plan out our design process, we knew there were two major aspects to focus on:

How Did We Achieve Our Goals?

Our audience is elusive, so we needed to fully inhabit their world, aiming to make the complex simple. Given the intricacies of modern repositories, we had to distill and iterate incessantly, to transform challenging concepts into a simple user experience — especially in a web app like Scholar, that must cover a lot of situations in simplified ways.

Through this iterative process, we confirmed our assumptions: by getting the analysis and discovery right, complexity became malleable and easy to handle in bite-sized iterations that accumulated into major features of the product. Here's some examples of how we, aligned with our principles, achieved our goals.

Gathering Insights

To navigate the niche landscape of our product, it was crucial for us to lean on expert recommendations and insights from key people in the field, through a series of Subject Matter Expert interviews:

We also benchmarked more than 10 scientific repositories, to analyze the current panorama. Some of our discoveries were:

  1. A clear focus on quick actions;
  2. Key informations displayed and structured in the same patterns (e.g research information has similar structure across 3+ repos);
  3. The act of importing/exporting publications is as paramount as creating new publications in a repository;
  4. There's opportunities to develop solutions that can analyze and resolve inconsistencies in publications, or detect duplicated work.

External Repository Integration

Researchers are sending their publications to other repos, and becoming faithful to their use, so how can we change that pattern?

With Scholar, we enabled researchers to link external repository accounts, allowing seamless publication imports to store past or external work. Users can also upload publications by importing reference management files, like Bibtex. We provide the option for Admins to customize the default repository list, so as to capture different sources other than ORCID, Scopus or DBLP.

Notification Central

In the past repos from faculties, users had no idea of what was going on, who edited their publications or the state of their reports…

But now, Scholar’s notification system provides tailored updates, keeping researchers updated on interactions with their work, while admins can track key activity across domains. Visual cues like color coding and icons simplify prioritization, to ensure fast responses.

Detection of Inconsistencies & Duplicates

In our user testing, we discovered that managing publications can be torture for researchers, especially when it involves manually checking for errors and duplicates.

Scholar detects and flags duplicates and inconsistencies in imported publications, identifying and suggesting identical content that can be merged manually or automatically. To ensure focus to these activities, users receive alerts to manage these suggestions directly in their Task area, keeping them on track of this work.

Impact

From its launch year (2020) until the time of writing, Scholar has managed to successfully capture and retain relevant scientific content produced across all faculties:

30155

Total Indexed Authors

110059

Total Indexed Publications

+7%

Publication Indexing Rate

If I Could Go Back…

I wanted to reflect on this product that I haven’t worked on since some years…to look, with fresh eyes, at features I'd improve if I could go back. One key improvement would be to enhance the duplicate management UI, so users quickly identify differences between similar publications. At the time, the idea of browsing through duplicates as if you’re flipping papers in your hand seemed clever, but through user testing, we saw users facing mental strain to memorize the differences between different duplicates.

With better judgement, and a much more calloused brain, I would implement a “diffs” feature to highlight differences directly on the interface, kind of like code review UIs! This fix would display variations between duplicates, sparing researchers from mentally tracking details and making the correction process much smoother.

Team Shoutouts

Thank you to the supertalented Tiago Alexandre Lopes, for teaching me everything I know about Product Design — I don’t know what would have been my career if we didn’t cross paths. You know how some people just have it? When it comes to designing products, Tiago has it, and more.

Big thanks also to Catarina Cepeda, Rita Severo and João Reis for building the foundations of Scholar, and to Tiago Machado for believing in my potential and giving me my first break in the industry!