What we do
Today, we generate more data than ever before. From restaurant reviews to turnstyle traffic, we create, log, and store electronic traces of our thoughts and behavior. We call these types of data BONDS -- big or naturally occurring data sets. While companies can use these data to anticipate market trends, researchers can sift through them to uncover meaningful insights into human cognition, behavior, and emotion.
Our goal at Data on the Mind is to bridge the gaps between BONDS and cognitive science: the imagination gap (or the difficulty of seeing a cognitive question in BONDS), the skills gap (or the difficulty of working with large and complex datasets), and the culture gap (or the difficulty in getting researchers and institutions to embrace BONDS). Currently, we accomplish this by providing lists of data resources with out-of-the-box cognitive potential, tools and tutorials to help along the way, and interviews with pioneering researchers to provide inspiration.
For more on our motivation, see our recent open-access article in Behavior Research Methods, "Finding the traces of behavior and cognitive process in big data and naturally occurring datasets (Paxton & Griffiths, in press).
Bridging the imagination gap: Data resources
With the widespread coverage about the power of BONDS, more and more researchers are expressing interest in tapping into these resources. One of the first hurdles that these researchers encounter is trying to find suitable data. Where are these troves of data, who can access them, and -- more importantly -- how can they be interpreted as remotely cognitive? Our data resources are aimed at answering those three questions.
We categorize the data we find into one of three types:
- Datasets: These are individual chunks of data about a single topic. Datasets may include only a single type of data or a range of multimodal data. Data collection for a single dataset may occur over any length of time, from days to decades. The defining factor for these data are that they are from or about a single source.
- Repositories: These are resources that encompass multiple individual datasets. Repositories may be open to all researchers or curated by a sponsoring institution. The datasets included may range wildly according to the missions of the specific repository. For researchers interested in making their data available to others, repositories are an excellent place to start.
- Lists: These are resources that can include both datasets and repositories. (Our data resource pages, for instance, would be classified as lists under our taxonomy.) Lists vary in the amount of information provided about the data they mention, but to be included here, they must include information about how to find or access the data. Lists may host the data that they mention, although they often don't.
Each of these data resources are tagged with a selection of fields or questions that might be interested in these data. These suggestions range from specific research areas (like categorization or language) to overarching fields (like health psychology or law), but -- with a little creativity -- we imagine many of these datasets shed light on even more fields and questions than we have identified. In addition to browsing the data by resource type and name, you can also browse through a complete list of our "applicable field" tags.
To start finding your next dataset, head over to our Find Data page.
Bridging the skills gap: Tools and tutorials
Decades of hardware and software advances are making it easier for cognitive scientists to make sense of complex data, but these can seem like a mystery to those who are just discovering BONDS. Complementing the data resources, our list of tools and tutorials can help researchers take the next steps after finding their data -- from statistical analysis to best practices writing for reproducible studies.
We categorize the resources we find into one of three types:
- Tutorials: These are direct instructions or walk-throughs. Tutorials may take a number of forms, including text, videos, and slideshows.
- Lists: These are aggregations of tutorials and other helpful information for a given methods topic.
- Tools: These provide a pointer to a useful tool. The level of documentation across tools will vary.
Where appropriate, we tag each with the applicable programming language(s). In addition to browsing all tools and tutorials, you can explore all tools and tutorials related to a specific language.
To discover the next big tool to help your research, head over to our Find Tools and Tutorials page.
Bridging the culture gap: Featured Projects
Pointing cognitive scientists to the data and tools to help them grow their research is an important step, but researchers who are new to this perspective may still be left wondering how to get started, how to deal with the new challenges of big data, and how to take full advantage of these unprecedented opportunities. To prime the imagination pump, then, we interview researchers who are actively working at the intersection of cognitive science and BONDS in a series that we call Featured Projects. We show you how big data and naturally occurring datasets can lead to new insights into human behavior and cognition by asking questions about how pioneers of this work find, work with, and interpret their data with important theoretical questions in mind. We reach out to researchers across disciplines and experience levels to ask them about their experiences to provide inspiration, prime creativity, and give useful advice on every stage of the scientific process.
To get inspired about how you can move from theory-driven question to big-data research, head over to our Featured Projects page.
Why it matters
Cognitive science faces a changing landscape. We have unprecedented opportunities to explore real-world behavior – and at the same time, we face new questions about reproducibility and open science. Successful researchers need to know how to take advantage of these new abilities while being mindful of the responsibilities that come with them.
We see our mission as highly related to issues of open science and reproducibility. By digging into troves of real-world data, researchers can work together to mine these data to better understand naturalistic behavior and cognition, and the data behind any resulting publications would be openly available for others to explore and reproduce. The open science movement takes data out of individual labs and introduces it to the wider scientific public, increasing transparency and credibility. In addition to curating lists of data and methods resources for this purpose, we intend to eventually assemble lists of resources and best practices to facilitate open science and reproducibility efforts.
Help us grow
We invite you to help us improve. If you have suggestions for how we can better meet our goals or for new Featured Projects, feel free to use our general suggestion form. If you know about resources that would be helpful to highlight, please suggest a data resource or a tool or tutorial. (Unfortunately, we are not currently hosting data, tools, or tutorials, so all suggestions must be available elsewhere.)