data science experts remotasks

data science experts remotasks

The Role of Data Science Experts in Remote Task Platforms

1. Introduction to Remote Task Platforms

Remote task platforms provide an option for managers to crowdsource tasks to a network of workers in fields such as design, software engineering, legal services, creative writing, marketing, and human sciences. Crowdworkers in the platform decide how much time to allocate to the job on their own, and the system pays according to the amount of work done. The community-based system distinguishes itself from traditional online labor platforms by allowing workers, as a community, to control some aspects of the platform. In comparison, for traditional online labor platforms, managers delegate tasks to workers. The potential for high turnover and the challenge of quality management resulting from employees not being physically present are barriers for managers to utilize remote workers. However, the existing research on remote task platforms has still primarily focused on how to improve the performance of crowdworkers. There is no research to date on the creation, execution, and appropriation of value on remote task platforms.

Although there is extensive literature studying freelancers or remote workers, the growing trend in employing specific employees in remote task platforms implies that the existing research primarily focuses on rather diversified workers. The current literature’s focus on workers implies that data from isolated individuals who work remotely are utilized. Therefore, we know much less about how crowdworkers interact with one another, communicate, share ideas and solutions, etc. within remote task platforms to overcome problems and improve job performance. Existing studies have appraised potential crowdworkers on remote task platforms, but creating, executing, managing, and adopting value on remote task platforms is not discussed in the literature. Unlike traditional online labor platforms based on the client-contractor model, remote task platforms attract FYBs (find your boss) or FXFs (flex your frequency) who are willing to spend time finding tasks or manage their time slots in the community and do the tasks when they are convenient.

2. Significance of Data Science in Remote Task Platforms

Remote task platforms (RTP) like Amazon Mechanical Turk have brought “human computation” to demand economies of scale. Human computation requires not only tools to distribute, track, and combine work distributed to communities of people but also tools to recruit domain experts who understand the results, design, refine, manage, evaluate, and assess the reliability of tasks, and understand stories they unlock. Domain experts are needed to transform all these tools from generic ones into task-specific ones, essential to extract higher levels of expertise from the workers. To support these sources of expertise, a new type of expert is required: the data science experts who are experienced in RTP practices. In this paper, we discuss the significance of data science experts in designing and solving remote task platforms as both people for building the platform and people using the platform.

Data-centric scientific discovery helps create a network that senses, thinks, and acts. Data Science is becoming an important area of research and development that supports the whole process of creating knowledge from massive data. The current development and application of Data Science can be scaled up by encouraging new types and levels of digital data collection and data-driven tasks. Examples of typical tasks on which Data Science tools give us game-changer capacities include image and culture recognition, style transfer, audio-visual translation, and language processing. Indeed, state-of-the-art tools are no longer designed to accomplish pre-assigned sub-tasks but are engaged in tutored learning. A promising, in-development tool is the HUB-C3PO model, an architecture that prunes language models to learn only once the filters of a small specialized model set. In research that aims to create ever-learning software, designing and solving human tasks will be essential.

3. Challenges and Opportunities for Data Science Experts

The advent of internet marketplaces has created opportunities for data scientists to work in or supplement their standard wage with platform work. Their roles have been diversified. Just as businesses are choosing among terms like “consulting”, “expert”, “experienced expert”, “services”, and “tasks” to view their operatives, potential employees are asked to decide on whether they should describe themselves as freelancers or consultants.

In the case of freelancers, they do not necessarily have to come from the high end of any profession such as engineers or data scientists. They can also come from the low end of the same professions such as online store data entry personnel engaged in creating and updating virtual stores for trading prints, design projects, hard-to-find antiques and artwork, etc.

Opportunities for the temporary workforce in various jobs conducted online can enable the occupation of data science among the contributions of online labor markets to employment patterns. Although the data collection interval is limited by design, we observed that online labor platforms tested have substantial talent in the data science task category.

Data science is often mislabeled as a subset of artificial intelligence, but while a data scientist’s work output feeds into an artificial intelligence process such as machine learning, that is not a required succession. Data Science is a big tent subject that covers everything from creating and manipulating data to analyzing and reporting on research data and making recommendations on how that data should be used, often while also creating and advertising training on non-data topics.

4. Best Practices and Strategies for Data Science Experts in Remote Task Platforms

One advantage of participating in remote-task platforms is that users can contribute their expertise in data science to various industry problems. In this section, we focus on the professional role of “data science experts” engaging in remote-task platforms. We address best practices and problems observed in the data annotation service and provide a few remedies to the problems.

First, it is difficult to discern expertise solely based on educational and professional experiences. Simple crowdsourcing platforms that offer easy access mask impressionable effectiveness. Therefore, asking a new employee to complete a few trial tasks before fully evaluating quality and speed is not feasible. However, one differentiator between people via a simple observation could be a degree in a relevant subject (e.g. professional or volunteer, or at least part-time job). This obvious trait represents an important signal for data consumers to recruit experts using remote-task platforms.

5. Future Trends and Innovations

Several potential future trends and innovations in crowdsourcing have been identified by different researchers in the literature so far. In order to round off this study, we introduce and discuss each of these trends, as well as any related subsets, under individual albeit interconnected categories. In doing so, we also illustrate the positioning of the open problems on the crowdsourcing map, as well as derive some additional themed areas for future research.

• Better Task Design and Task Sourcing: Execution of mechanistic, human intelligence tasks made widely available on today’s remote-task platforms is often challenging due to issues in task design. This results in low response rates and therefore poor access to data in a wide range of scientific and business domains. We believe this problem is often due to a lack of ability in developing high-quality tasking instructions with explicit examples and a lack of task sourcing expertise for knowing which tasks to create and post. These issues can be avoided by making transparent the design of successful tasking instructions and accelerating the development of new tasking instructions by either controversializing existing instructions or by gamifying and monetizing their redesign process.