Data Valuation Platform

The data deluge resulting from the technological innovation that accompanies our day-to-day lives has defined digital economies and created the proper context for the emergence of data markets. With it a need to understand what is the value of data.

Data value is hard to define, because the same data is valued differently by various businesses (e.g., a set of GPS traces has different uses for a ride hailing company, a public administration, a retailer) or even by various departments within the same business (Top management vs. Technical vs. Mid-management).

Type

R&D

Big Data

Role

Product Designer

Tools

Figma

Miro

Slack

Zoom

Google Suite

Contribution

Surveys

Ideation

Paper wireframes

High-fi prototypes

Duration

4 months

Define

Challenge

Design an app that supports the collection of all the information necessary for the data valuation process. Some initial considerations:

Users with different roles perceive data value differently. Data value assessment has to adapt to different roles (perspectives).

The value of data is a complex concept and so is the process for establishing it. Users are expected to have a certain level of expertise.

Independent value assessments aggregate under a global data value. Users need to look at the value of data from both individual and global perspectives.

Context

The app is an extension of the Horizon 2020 Safe-DEED project. Its goal was to develop a data valuation platform, initially with a very basic UI. Following the positive feedback received for version 1 and with the perspectives of continuing the technological R&D, the necessity for improved UX and UI became appeared.

Research

Market Research

Hardly affordable for everyone.
In-house presence.
Use complex methodologies.
Lenghty processes.

Excel-based macros.
Unappealing UX and UI.
Requires a consultant.

Different methodology.
Focused only on data quality.

User Surveys

An online survey was sent to professionals from organisations with a data centred activity (involving the production or consumption of data). Most of these companies are Spanish-based and have a global reach; the rest are also based in Europe.

The survey contained 15 questions and sought to shed a light on the following:

How willing is a company to buy/sell data?
What kind of data is usually bought/sold?
What are the data characteristics that are important to organisations that buy/sell data?
What kind of output (reporting) are users expecting from a data valuation platform?

Is your organisation buying or selling data?

Which industry is your organisation active in?

Rate the importance of each of the following data properties

Rate the importance of the following types of reports.

Main findings

Buying Data

Companies are willing to buy external data.

Companies connect data value to data quality.

Selling Data

Companies are reluctant about selling their data.

Main reasons: legal, security, commercial.

Desired Output

Dataset overview.

Data quality assessment.

Economic value.

Exploitability assessment.

Analysis

User personas

Pain points

Compromise between process length and the necessity to gather exhaustive information.

Compromise between reporting a unique value and the need to know the little pieces that compose it.

Data quality assessment requires user input for a subset of data columns. Users need a fast filter-and-choose functionality.

Goal statement

The Data Valuation Platform will enable users to conduct valuations of their data sets, which will affect professionals from data-enabled businesses by allowing them to assess the quality and usability of data sets in a context that they define.

Design

User flow diagrams

The decision blocks (purple) are represented in dotted lines because they depend on the internal state of the application, and not directly on a user’s decision.

The main flow can progress towards any of the incomplete surveys. When all 3 surveys are complete, it displays the Data Value Scorecard.

Site map

The sitemap uses a colour code (see upper left corner) to describe the different user roles and their interaction with the different parts of the app.

Ideation

Working together with the Lead Researcher tasked with the development of the platform, we tried to improve the results of v1, particularly around the Pain Points identified above. I proposed sketches with solutions to these particular problems, which we refined over several brainstorming sessions.

Concept validation

The resulting prototypes were presented to various potential users (project partners, companies interested in the product), who had previously been exposed to version 1. The two solutions were presented comparatively, and feedback was collected.

Solution

Wrap Up

Lessons learned

Involve designers early

This was a valuable lesson both for the original project team, as well as for me as a designer. While the team was trying to elicit requirements using a software engineering approach, they realised the shortcomings of not having a more user-centred methodology after their first release. From my side as a (junior) designer freshly on board of a complex on-going project, it was very challenging to try to adjust their results to a design thinking approach.

Be ready for changes in scale and scope

The project started as a small, contained component that would lead to a proof of concept. I had to be able to jump on board and adapt to a change of scope, following the initial positive feedback.

Keep it flexible

From a business perspective, there are several ways in which the product can be monetised. It will also be included as a component in a new project, which will see slight adaptations to its functionality. I tried to design both the experience and the interface such that they are as adjustable as possible to these scenarios.

Know the application domain

The project itself is pushing the state-of-the-art in its field, and brings together concepts such as data quality, data privacy, machine learning etc. It was fascinating to try to get some (very basic!) understanding of it all and see what big data R&D teams are working on these days.

Accessibility

This is something that needs to be addressed early on in the next iteration, primarily in terms of visual accessibility (colour scheme, scores presentations, charts).

Tutorials

The users are dealing with a long process of data assessment. Our first tests show that once they get used to it, they become very proficient with using the platform. However, the onboarding can still be difficult. I plan to address this by developing a set of visual tutorials, aimed at the 3 different roles, and focused on explaining the basics for each of the 3 main user flows.

Tooltips

While we know that the users of this product are data-savvy, they are still exposed to specialised terms, that need either to be defined (e.g., enterprise data, personal data) or further clarified (e.g., types of data licenses, the GDPR principles, what qualifies as missing data). I am trying to solve this by experimenting with tooltips positioned in various places of the interface.