Customer Login  |  

One of the corollaries to the big data meme is that the world is woefully short of data science professionals. This observation requires some context, however, because the definition of what a data scientist _is_ is a matter of some debate. Some define data science as statistical analysis while others look at data as “information” or “content” and see the science of data as more related to the creation, management, and curation of content.

In terms of the education required for a career in data science, students’ approaches usually depend on their own definition of the data science career path. In other words, statisticians prefer to take more math and economics classes than those who prepare themselves for the content business by taking English and other liberal arts courses.

Of all the graduate programs available to future data scientists, however, the one that is the most intriguing is information science. There are several outstanding graduate schools of “information”—Information Evolution is a proud supporter of the University of Texas’s School of Information—all of which were once called library schools. (N.b.: The SIIA Content Division has graciously allowed UT information school graduates to attend their events in exchange for volunteer time. It seems that they, too, recognize these smart young folks as an important part of the future of our industry.)

These programs continue to aggressively include more technical classes as the world of information becomes more and more electronic. Human-computer interfaces, for instance, are the essential gateways to information, so information school graduates leave these programs with a deep understanding of the design and function of these interfaces. Similarly, preserving digital information is an emerging area of concern for all purveyors of online information. Information schools are also one of the few places in academia analyzing how electronic information can be truly preserved for the analysis of posterity.

Other critical issues related to the business of information include the ownership and dissemination of information, limitations on individual or corporate “rights” to privacy, and use of user-generated and crowd-sourced information. No other academic discipline produces graduates so uniquely qualified to understand the context of information (who uses it, why they use it, etc.), which is absolutely essential to designing and running these services, the way that information schools do. This is what makes them such an important resource both to our public institutions and, increasingly, to the companies in the business of selling information.

{ 0 comments }

posted by Shyamali Ghosh on June 18, 2013

by Kevin Dodds

Getting the right workers is critically important for most managed crowdsourcing projects. Success depends on a pool of workers who understand and can do the tasks involved. So, while it’s certainly possible to open a project up to the entire crowd and then start blocking poor performers along the way, gradually thinning the worker pool down to the best of the best, it’s important to recognize that qualifying crowd workers for a campaign is far preferable to blocking them.

Putting a “block” on a worker has negative effects on his or her account and ability to work, so blocking a worker can be seen as an aggressive act, comparable to the firing of an employee. There’s bad feeling on both sides, and the worker will have trouble picking up a new job, even one they’ve already shown they’re suited for. This can also lead to bad social chatter about your firm from the crowdworker community.

Conversely, qualifying (and disqualifying) workers is a more equitable way of putting together an excellent worker pool that avoids negatively affecting workers. By qualifying workers—making sure that they can do tasks with known answers before allowing them to work on live records—it’s possible to pinpoint those people who are not well-suited to a certain task as well as those who are exceptionally adept.

It’s also important to communicate with workers who seem to have trouble hitting gold standards or qualifications. For example, when a worker with a very high gold accuracy rating suddenly hits zero on a specific campaign run, something is clearly wrong. A word from the project manager often helps the crowd worker become aware of a simple copy and paste error and allows him or her to correct the mistake. A manager could, of course, simply look at the numbers and boot the offender right out of the pool. In many cases, though, little errors are made by capable, useful workers who can benefit the project and deserve a second chance.

Qualifying crowd workers, means, essentially, treating them as you would an in-house employee: making sure that they’re able to do the work and monitoring them to ensure they continue to meet the standard you’ve set. This leads to higher overall quality and the likelihood that these highly qualified resources will participate in future projects.

{ 0 comments }

posted by Shyamali Ghosh on June 18, 2013